feat(import): update rows on primary-key conflicts

Closes #10014
This commit is contained in:
t8y2 committed 2026-09-27 05:54:32 +00:00
1 parent d5dcdabd22
commit 5f8bcf74a5
12 files changed
+758 -110

No files matched your search

@@ -140,7 +140,7 @@ const titleRow = ref(1);
const dataStartRow = ref(2); const dataStartRow = ref(2);
const lastDataRow = ref(0); const lastDataRow = ref(0);
const trimValues = ref(false); const trimValues = ref(false);
const skipDuplicateRows = ref(false); const conflictPolicy = ref<api.TableImportConflictPolicy>("error");
const emptyStringAsNull = ref(defaultTableImportEmptyStringAsNull(sourceFormat.value)); const emptyStringAsNull = ref(defaultTableImportEmptyStringAsNull(sourceFormat.value));
const selectedSheet = ref(""); const selectedSheet = ref("");
const jsonShape = ref<api.TableImportJsonShape>("auto"); const jsonShape = ref<api.TableImportJsonShape>("auto");
@@ -176,10 +176,10 @@ const wizardSteps: Array<{ value: TableImportWizardStep; labelKey: string }> = [
const selectedConnection = computed(() => (props.prefillConnectionId ? store.getConfig(props.prefillConnectionId) : undefined)); const selectedConnection = computed(() => (props.prefillConnectionId ? store.getConfig(props.prefillConnectionId) : undefined));
const structureDatabaseType = computed(() => tableStructureDatabaseTypeForConnection(selectedConnection.value)); const structureDatabaseType = computed(() => tableStructureDatabaseTypeForConnection(selectedConnection.value));
// Mirrors the dialect list in skip-generator dispatch (transfer.rs): only these // Mirrors the conflict SQL dispatch in transfer.rs. Other dialects must keep
// engines turn the skip-duplicates option into conflict-handling INSERTs. // ordinary INSERT/error behavior instead of approximating an upsert.
const SKIP_DUPLICATE_ROWS_DATABASE_TYPES = new Set<DatabaseType>(["postgres", "kingbase", "opengauss", "sqlite", "cloudflare-d1", "duckdb", "mysql", "doris", "starrocks"]); const IMPORT_CONFLICT_DATABASE_TYPES = new Set<DatabaseType>(["postgres", "kingbase", "opengauss", "sqlite", "cloudflare-d1", "duckdb", "mysql", "doris", "starrocks"]);
const supportsSkipDuplicateRows = computed(() => structureDatabaseType.value !== undefined && SKIP_DUPLICATE_ROWS_DATABASE_TYPES.has(structureDatabaseType.value)); const supportsImportConflictPolicy = computed(() => structureDatabaseType.value !== undefined && IMPORT_CONFLICT_DATABASE_TYPES.has(structureDatabaseType.value));
const targetSchema = computed(() => metadataSchemaForConnection(selectedConnection.value, props.prefillDatabase || "", props.prefillSchema)); const targetSchema = computed(() => metadataSchemaForConnection(selectedConnection.value, props.prefillDatabase || "", props.prefillSchema));
const dataTypeOptions = computed(() => mergeDataTypeOptions(dynamicDataTypeOptions.value, getDataTypeOptions(structureDatabaseType.value), Object.values(columnDataTypes.value))); const dataTypeOptions = computed(() => mergeDataTypeOptions(dynamicDataTypeOptions.value, getDataTypeOptions(structureDatabaseType.value), Object.values(columnDataTypes.value)));
const hasExistingTarget = computed(() => !!props.prefillTable || loadingExistingTables.value || existingTableNames.value.length > 0); const hasExistingTarget = computed(() => !!props.prefillTable || loadingExistingTables.value || existingTableNames.value.length > 0);
@@ -202,6 +202,20 @@ const mappedColumns = computed<api.TableImportColumnMapping[]>(() => {
}); });
const mappedCount = computed(() => mappedColumns.value.length); const mappedCount = computed(() => mappedColumns.value.length);
const mappingValidation = computed(() => validateImportMappings(mappedColumns.value)); const mappingValidation = computed(() => validateImportMappings(mappedColumns.value));
const primaryKeyColumns = computed(() => targetColumns.value.filter((column) => column.is_primary_key));
const updateExistingUnavailableReason = computed(() => {
if (targetMode.value !== "existing") return t("tableImport.updateExistingRequiresExistingTable");
if (!supportsImportConflictPolicy.value) return t("tableImport.updateExistingUnsupportedDialect");
if (!existingTargetMetadataReady.value || primaryKeyColumns.value.length === 0) return t("tableImport.updateExistingRequiresPrimaryKey");
const mappedTargets = mappedColumns.value.map((mapping) => mapping.targetColumn.toLowerCase());
const missingKeys = primaryKeyColumns.value.filter((column) => !mappedTargets.includes(column.name.toLowerCase()));
if (missingKeys.length > 0) return t("tableImport.updateExistingRequiresMappedPrimaryKey", { columns: missingKeys.map((column) => column.name).join(", ") });
const primaryKeyNames = primaryKeyColumns.value.map((column) => column.name.toLowerCase());
if (!mappedTargets.some((column) => !primaryKeyNames.includes(column))) return t("tableImport.updateExistingRequiresMappedValue");
return "";
});
const canUpdateExistingRows = computed(() => updateExistingUnavailableReason.value === "");
const conflictPolicyError = computed(() => (conflictPolicy.value === "updateExisting" ? updateExistingUnavailableReason.value : ""));
const requiredUnmappedColumns = computed(() => const requiredUnmappedColumns = computed(() =>
requiredImportTargetColumns( requiredImportTargetColumns(
targetColumns.value, targetColumns.value,
@@ -215,7 +229,7 @@ const batchTargetNamesValid = computed(() => {
return tableNames.length > 0 && tableNames.every(Boolean) && new Set(tableNames).size === tableNames.length; return tableNames.length > 0 && tableNames.every(Boolean) && new Set(tableNames).size === tableNames.length;
}); });
const canImport = computed(() => { const canImport = computed(() => {
if (running.value || !props.prefillConnectionId || !existingTargetMetadataReady.value) return false; if (running.value || !props.prefillConnectionId || !existingTargetMetadataReady.value || conflictPolicyError.value) return false;
if (!isBatchImport.value) return !!preview.value && !!targetTableName.value && mappingValidation.value.valid; if (!isBatchImport.value) return !!preview.value && !!targetTableName.value && mappingValidation.value.valid;
if (!batchTargetNamesValid.value) return false; if (!batchTargetNamesValid.value) return false;
return selectedBatchTasks.value.every((task) => { return selectedBatchTasks.value.every((task) => {
@@ -231,7 +245,7 @@ const canGoNext = computed(() => {
if (wizardStep.value === "mapping") return canImport.value; if (wizardStep.value === "mapping") return canImport.value;
} }
if (wizardStep.value === "options") return !!preview.value && !!targetTableName.value && existingTargetMetadataReady.value; if (wizardStep.value === "options") return !!preview.value && !!targetTableName.value && existingTargetMetadataReady.value;
if (wizardStep.value === "mapping") return existingTargetMetadataReady.value && mappingValidation.value.valid; if (wizardStep.value === "mapping") return existingTargetMetadataReady.value && mappingValidation.value.valid && !conflictPolicyError.value;
return false; return false;
}); });
const rawProgressPercent = computed(() => tableImportProgressPercent(progress.value)); const rawProgressPercent = computed(() => tableImportProgressPercent(progress.value));
@@ -384,7 +398,7 @@ function resetState() {
dataStartRow.value = 2; dataStartRow.value = 2;
lastDataRow.value = 0; lastDataRow.value = 0;
trimValues.value = false; trimValues.value = false;
skipDuplicateRows.value = false; conflictPolicy.value = "error";
emptyStringAsNull.value = defaultTableImportEmptyStringAsNull(sourceFormat.value); emptyStringAsNull.value = defaultTableImportEmptyStringAsNull(sourceFormat.value);
selectedSheet.value = ""; selectedSheet.value = "";
jsonShape.value = "auto"; jsonShape.value = "auto";
@@ -851,7 +865,7 @@ function canOpenStep(step: TableImportWizardStep) {
if (step === "review") return canImport.value; if (step === "review") return canImport.value;
} }
if (step === "mapping") return !!preview.value && existingTargetMetadataReady.value; if (step === "mapping") return !!preview.value && existingTargetMetadataReady.value;
if (step === "review") return !!preview.value && existingTargetMetadataReady.value && mappingValidation.value.valid; if (step === "review") return !!preview.value && existingTargetMetadataReady.value && mappingValidation.value.valid && !conflictPolicyError.value;
return false; return false;
} }
@@ -941,7 +955,8 @@ async function startImport() {
mode: targetMode.value === "create" ? "append" : importMode.value, mode: targetMode.value === "create" ? "append" : importMode.value,
createTable: targetMode.value === "create", createTable: targetMode.value === "create",
batchSize: Math.max(1, Number(batchSize.value) || 500), batchSize: Math.max(1, Number(batchSize.value) || 500),
skipDuplicateRows: supportsSkipDuplicateRows.value && skipDuplicateRows.value, conflictPolicy: conflictPolicy.value,
skipDuplicateRows: conflictPolicy.value === "skip",
dateTimeFormat: settingsStore.editorSettings.globalDateTimeImportFormat || undefined, dateTimeFormat: settingsStore.editorSettings.globalDateTimeImportFormat || undefined,
preparedSource: preparedImportSource(currentPreview), preparedSource: preparedImportSource(currentPreview),
}, },
@@ -1031,7 +1046,8 @@ async function startBatchImport() {
mode: "append", mode: "append",
createTable: true, createTable: true,
batchSize: Math.max(1, Number(batchSize.value) || 500), batchSize: Math.max(1, Number(batchSize.value) || 500),
skipDuplicateRows: supportsSkipDuplicateRows.value && skipDuplicateRows.value, conflictPolicy: conflictPolicy.value,
skipDuplicateRows: conflictPolicy.value === "skip",
dateTimeFormat: settingsStore.editorSettings.globalDateTimeImportFormat || undefined, dateTimeFormat: settingsStore.editorSettings.globalDateTimeImportFormat || undefined,
preparedSource: preparedImportSource(task.preview), preparedSource: preparedImportSource(task.preview),
retainSource: true, retainSource: true,
@@ -1198,13 +1214,17 @@ watch(targetMode, (mode) => {
loadedTargetTableName.value = ""; loadedTargetTableName.value = "";
loadingTarget.value = false; loadingTarget.value = false;
importMode.value = "append"; importMode.value = "append";
conflictPolicy.value = "error";
applyAutoMapping(); applyAutoMapping();
applySuggestedColumnDataTypes(); applySuggestedColumnDataTypes();
void loadDataTypeOptions(); void loadDataTypeOptions();
} }
}); });
watch(selectedExistingTable, () => { watch(selectedExistingTable, () => {
if (targetMode.value === "existing") void loadTargetColumns(); if (targetMode.value === "existing") {
conflictPolicy.value = "error";
void loadTargetColumns();
}
}); });
watch(rawProgressPercent, (percent) => { watch(rawProgressPercent, (percent) => {
@@ -1422,10 +1442,6 @@ watch(rawProgressPercent, (percent) => {
<input v-model="emptyStringAsNull" type="checkbox" class="h-3.5 w-3.5 accent-primary" /> <input v-model="emptyStringAsNull" type="checkbox" class="h-3.5 w-3.5 accent-primary" />
{{ t("tableImport.emptyStringAsNull") }} {{ t("tableImport.emptyStringAsNull") }}
</label> </label>
<label v-if="supportsSkipDuplicateRows" class="flex items-center gap-2 text-xs">
<input v-model="skipDuplicateRows" type="checkbox" class="h-3.5 w-3.5 accent-primary" />
{{ t("tableImport.skipDuplicateRows") }}
</label>
</div> </div>
<div v-else-if="sourceFormat === 'sql'" class="grid grid-cols-5 gap-3 rounded-md border p-3"> <div v-else-if="sourceFormat === 'sql'" class="grid grid-cols-5 gap-3 rounded-md border p-3">
@@ -1490,6 +1506,20 @@ watch(rawProgressPercent, (percent) => {
</div> </div>
</div> </div>
<div v-if="supportsImportConflictPolicy" class="space-y-1.5 rounded-md border p-3">
<Label for="table-import-conflict-policy" class="text-xs">{{ t("tableImport.conflictPolicy") }}</Label>
<select id="table-import-conflict-policy" v-model="conflictPolicy" data-testid="table-import-conflict-policy" class="flex h-8 w-full rounded-md border border-input bg-background px-2 text-xs">
<option value="error">{{ t("tableImport.conflictError") }}</option>
<option value="skip">{{ t("tableImport.skipDuplicateRows") }}</option>
<option v-if="targetMode === 'existing'" value="updateExisting" :disabled="!canUpdateExistingRows">
{{ t("tableImport.updateExistingRows") }}
</option>
</select>
<p v-if="targetMode === 'existing' && !canUpdateExistingRows" class="text-[11px] text-muted-foreground">
{{ updateExistingUnavailableReason }}
</p>
</div>
<div class="flex items-center gap-2"> <div class="flex items-center gap-2">
<Button size="sm" :disabled="!selectedSource || loadingPreview" @click="loadPreview()"> <Button size="sm" :disabled="!selectedSource || loadingPreview" @click="loadPreview()">
<Loader2 v-if="loadingPreview" class="mr-1.5 h-3.5 w-3.5 animate-spin" /> <Loader2 v-if="loadingPreview" class="mr-1.5 h-3.5 w-3.5 animate-spin" />
@@ -1597,6 +1627,9 @@ watch(rawProgressPercent, (percent) => {
<div v-if="mappingValidation.errors.length" class="rounded-md border border-destructive/30 bg-destructive/10 px-3 py-2 text-xs text-destructive"> <div v-if="mappingValidation.errors.length" class="rounded-md border border-destructive/30 bg-destructive/10 px-3 py-2 text-xs text-destructive">
{{ mappingValidation.errors.join("; ") }} {{ mappingValidation.errors.join("; ") }}
</div> </div>
<div v-else-if="conflictPolicyError" class="rounded-md border border-destructive/30 bg-destructive/10 px-3 py-2 text-xs text-destructive">
{{ conflictPolicyError }}
</div>
<div v-else-if="requiredUnmappedColumns.length" class="flex items-start gap-2 rounded-md border border-amber-300 bg-amber-50 px-3 py-2 text-xs text-amber-800 dark:bg-amber-950/20 dark:text-amber-300"> <div v-else-if="requiredUnmappedColumns.length" class="flex items-start gap-2 rounded-md border border-amber-300 bg-amber-50 px-3 py-2 text-xs text-amber-800 dark:bg-amber-950/20 dark:text-amber-300">
<AlertTriangle class="mt-0.5 h-3.5 w-3.5 shrink-0" /> <AlertTriangle class="mt-0.5 h-3.5 w-3.5 shrink-0" />
<span>{{ t("tableImport.requiredUnmapped", { columns: requiredUnmappedColumns.join(", ") }) }}</span> <span>{{ t("tableImport.requiredUnmapped", { columns: requiredUnmappedColumns.join(", ") }) }}</span>
@@ -294,6 +294,7 @@ describe("TableImportDialog batch selection", () => {
await flushAsyncUpdates(); await flushAsyncUpdates();
expect(document.body.textContent).toContain("Trim values"); expect(document.body.textContent).toContain("Trim values");
expect(document.body.textContent).not.toContain("Skip duplicate rows"); expect(document.body.textContent).not.toContain("Skip duplicate rows");
expect(document.body.textContent).not.toContain("Update the existing row");
}); });
it("offers the skip-duplicates option for dialects with conflict handling", async () => { it("offers the skip-duplicates option for dialects with conflict handling", async () => {
@@ -304,6 +305,25 @@ describe("TableImportDialog batch selection", () => {
expect(document.body.textContent).toContain("Skip duplicate rows"); expect(document.body.textContent).toContain("Skip duplicate rows");
}); });
it("serializes the explicit skip policy with the legacy compatibility flag", async () => {
mocks.previewTableImportFile.mockImplementation(() => Promise.resolve(delimitedPreview()));
await mountDialog([new File(["id,name\n1,a"], "rows.csv")]);
const policySelect = document.body.querySelector<HTMLSelectElement>('[data-testid="table-import-conflict-policy"]');
expect(policySelect).toBeTruthy();
if (policySelect) {
policySelect.value = "skip";
policySelect.dispatchEvent(new Event("change", { bubbles: true }));
}
await flushAsyncUpdates();
await startImport();
expect(mocks.importTableFile).toHaveBeenCalledTimes(1);
expect(mocks.importTableFile.mock.calls[0]![0]).toMatchObject({
conflictPolicy: "skip",
skipDuplicateRows: true,
});
});
it("keeps every worksheet selected by default and imports them in workbook order", async () => { it("keeps every worksheet selected by default and imports them in workbook order", async () => {
await mountDialog(); await mountDialog();
for (const name of ["rows_First", "rows_Second", "rows_Third"]) expect(taskCheckbox(name).checked).toBe(true); for (const name of ["rows_First", "rows_Second", "rows_Third"]) expect(taskCheckbox(name).checked).toBe(true);
@@ -3,6 +3,7 @@
import { createApp, defineComponent, h, nextTick, type App } from "vue"; import { createApp, defineComponent, h, nextTick, type App } from "vue";
import { afterEach, beforeEach, describe, expect, it, vi } from "vitest"; import { afterEach, beforeEach, describe, expect, it, vi } from "vitest";
import i18n from "@/i18n"; import i18n from "@/i18n";
import type { TableImportProgress, TableImportRequest, TableImportSummary } from "@/lib/backend/api";
const mocks = vi.hoisted(() => ({ const mocks = vi.hoisted(() => ({
ensureConnected: vi.fn().mockResolvedValue(undefined), ensureConnected: vi.fn().mockResolvedValue(undefined),
@@ -11,8 +12,8 @@ const mocks = vi.hoisted(() => ({
{ name: "archived_target", table_type: "TABLE" }, { name: "archived_target", table_type: "TABLE" },
]), ]),
getColumns: vi.fn().mockResolvedValue([ getColumns: vi.fn().mockResolvedValue([
{ name: "id", data_type: "INTEGER", nullable: false }, { name: "id", data_type: "INTEGER", nullable: false, is_primary_key: true },
{ name: "name", data_type: "TEXT", nullable: true }, { name: "name", data_type: "TEXT", nullable: true, is_primary_key: false },
]), ]),
listDataTypes: vi.fn().mockResolvedValue(["INTEGER", "TEXT"]), listDataTypes: vi.fn().mockResolvedValue(["INTEGER", "TEXT"]),
previewTableImportFile: vi.fn().mockResolvedValue({ previewTableImportFile: vi.fn().mockResolvedValue({
@@ -26,6 +27,7 @@ const mocks = vi.hoisted(() => ({
sourceFingerprint: "rows-xlsx", sourceFingerprint: "rows-xlsx",
sheets: ["Data"], sheets: ["Data"],
}), }),
importTableFile: vi.fn<(request: TableImportRequest, onProgress: (progress: TableImportProgress) => void) => Promise<TableImportSummary>>(),
releaseTableImportSource: vi.fn().mockResolvedValue(undefined), releaseTableImportSource: vi.fn().mockResolvedValue(undefined),
})); }));
@@ -52,7 +54,7 @@ vi.mock("@/lib/backend/api", () => ({
listDataTypes: mocks.listDataTypes, listDataTypes: mocks.listDataTypes,
previewTableImportFile: mocks.previewTableImportFile, previewTableImportFile: mocks.previewTableImportFile,
releaseTableImportSource: mocks.releaseTableImportSource, releaseTableImportSource: mocks.releaseTableImportSource,
importTableFile: vi.fn(), importTableFile: mocks.importTableFile,
cancelTableImport: vi.fn(), cancelTableImport: vi.fn(),
})); }));
@@ -208,7 +210,12 @@ async function selectWorkbook() {
value: [new File(["test workbook"], "rows.xlsx", { type: "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet" })], value: [new File(["test workbook"], "rows.xlsx", { type: "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet" })],
}); });
fileInput?.dispatchEvent(new Event("change", { bubbles: true })); fileInput?.dispatchEvent(new Event("change", { bubbles: true }));
await vi.waitFor(() => expect(mocks.previewTableImportFile).toHaveBeenCalled()); await flushAsyncUpdates();
if (!mocks.previewTableImportFile.mock.calls.length) {
buttonContaining("Load Preview")?.click();
}
await vi.waitFor(() => expect(mocks.previewTableImportFile).toHaveBeenCalled(), { timeout: 3000 });
await flushAsyncUpdates();
} }
async function selectExistingTarget(tableName: string) { async function selectExistingTarget(tableName: string) {
@@ -232,11 +239,15 @@ beforeEach(() => {
{ name: "archived_target", table_type: "TABLE" }, { name: "archived_target", table_type: "TABLE" },
]); ]);
mocks.getColumns.mockReset().mockResolvedValue([ mocks.getColumns.mockReset().mockResolvedValue([
{ name: "id", data_type: "INTEGER", nullable: false }, { name: "id", data_type: "INTEGER", nullable: false, is_primary_key: true },
{ name: "name", data_type: "TEXT", nullable: true }, { name: "name", data_type: "TEXT", nullable: true, is_primary_key: false },
]); ]);
mocks.listDataTypes.mockClear(); mocks.listDataTypes.mockClear();
mocks.previewTableImportFile.mockClear(); mocks.previewTableImportFile.mockClear();
mocks.importTableFile.mockReset().mockImplementation(async (request, onProgress) => {
onProgress({ importId: request.importId, status: "done", rowsImported: 1, totalRows: 1, elapsedMs: 1 });
return { importId: request.importId, rowsImported: 1, totalRows: 1, elapsedMs: 1 };
});
mocks.releaseTableImportSource.mockClear(); mocks.releaseTableImportSource.mockClear();
}); });
@@ -346,4 +357,60 @@ describe("TableImportDialog existing targets", () => {
expect(mocks.listTables).not.toHaveBeenCalled(); expect(mocks.listTables).not.toHaveBeenCalled();
expect(mocks.getColumns).toHaveBeenCalledWith("connection-1", "main", "main", "existing_target"); expect(mocks.getColumns).toHaveBeenCalledWith("connection-1", "main", "main", "existing_target");
}); });
it("offers update only with primary-key metadata and serializes the explicit policy", async () => {
i18n.global.locale.value = "en";
await mountDialog({ schema: "main" });
await selectWorkbook();
buttonContaining("Existing table")?.click();
await flushAsyncUpdates();
await selectExistingTarget("existing_target");
await vi.waitFor(() => expect(buttonContaining("Next")?.disabled).toBe(false));
const policySelect = document.body.querySelector<HTMLSelectElement>('[data-testid="table-import-conflict-policy"]');
expect(policySelect).toBeTruthy();
const updateOption = policySelect?.querySelector<HTMLOptionElement>('option[value="updateExisting"]');
await vi.waitFor(() => expect(updateOption?.disabled, document.body.textContent || "").toBe(false));
if (policySelect) {
policySelect.value = "updateExisting";
policySelect.dispatchEvent(new Event("change", { bubbles: true }));
}
await flushAsyncUpdates();
buttonContaining("Next")?.click();
await flushAsyncUpdates();
buttonContaining("Next")?.click();
await flushAsyncUpdates();
buttonContaining("Start Import")?.click();
await flushAsyncUpdates();
expect(mocks.importTableFile).toHaveBeenCalledTimes(1);
expect(mocks.importTableFile.mock.calls[0]![0]).toMatchObject({
conflictPolicy: "updateExisting",
skipDuplicateRows: false,
mappings: [
{ sourceColumn: "id", targetColumn: "id" },
{ sourceColumn: "name", targetColumn: "name" },
],
});
});
it("disables update when target primary-key metadata is absent", async () => {
mocks.getColumns.mockResolvedValueOnce([
{ name: "id", data_type: "INTEGER", nullable: false, is_primary_key: false },
{ name: "name", data_type: "TEXT", nullable: true, is_primary_key: false },
]);
i18n.global.locale.value = "en";
await mountDialog({ schema: "main" });
await selectWorkbook();
buttonContaining("Existing table")?.click();
await flushAsyncUpdates();
await selectExistingTarget("existing_target");
await vi.waitFor(() => expect(mocks.getColumns).toHaveBeenCalled());
await flushAsyncUpdates();
const updateOption = document.body.querySelector<HTMLOptionElement>('option[value="updateExisting"]');
expect(updateOption?.disabled).toBe(true);
expect(document.body.textContent).toContain("requires primary-key metadata");
});
}); });
+8
View File
@@ -6856,7 +6856,15 @@ export default {
lastDataRow: "Last data row (0 for end)", lastDataRow: "Last data row (0 for end)",
trimValues: "Trim values", trimValues: "Trim values",
emptyStringAsNull: "Empty string as NULL", emptyStringAsNull: "Empty string as NULL",
conflictPolicy: "On primary-key conflict",
conflictError: "Stop with an error",
skipDuplicateRows: "Skip duplicate rows", skipDuplicateRows: "Skip duplicate rows",
updateExistingRows: "Update the existing row",
updateExistingRequiresExistingTable: "Updating conflicts requires an existing target table.",
updateExistingUnsupportedDialect: "Updating conflicts is not supported for this database.",
updateExistingRequiresPrimaryKey: "Updating conflicts requires primary-key metadata for the target table.",
updateExistingRequiresMappedPrimaryKey: "Map every primary-key column to update conflicts. Missing: {columns}",
updateExistingRequiresMappedValue: "Map at least one non-primary-key column to update conflicts.",
jsonShape: "JSON rows", jsonShape: "JSON rows",
jsonShapeAuto: "Auto-detect", jsonShapeAuto: "Auto-detect",
jsonShapeObjects: "Objects", jsonShapeObjects: "Objects",
+9
View File
@@ -6823,6 +6823,15 @@ export default withEnglishFallback({
lastDataRow: "最后数据行(0 表示末尾)", lastDataRow: "最后数据行(0 表示末尾)",
trimValues: "裁剪空白", trimValues: "裁剪空白",
emptyStringAsNull: "空字符串作为 NULL", emptyStringAsNull: "空字符串作为 NULL",
conflictPolicy: "主键冲突时",
conflictError: "停止并报错",
skipDuplicateRows: "跳过重复行",
updateExistingRows: "更新已有行",
updateExistingRequiresExistingTable: "冲突更新仅适用于已有目标表。",
updateExistingUnsupportedDialect: "当前数据库不支持冲突更新。",
updateExistingRequiresPrimaryKey: "冲突更新需要目标表的主键元数据。",
updateExistingRequiresMappedPrimaryKey: "冲突更新需要映射所有主键列,缺少:{columns}",
updateExistingRequiresMappedValue: "冲突更新需要至少映射一个非主键列。",
jsonShape: "JSON 行结构", jsonShape: "JSON 行结构",
jsonShapeAuto: "自动检测", jsonShapeAuto: "自动检测",
jsonShapeObjects: "对象", jsonShapeObjects: "对象",
+9
View File
@@ -5525,6 +5525,15 @@ export default withEnglishFallback({
hasHeader: "首列為標頭", hasHeader: "首列為標頭",
trimValues: "裁剪空白", trimValues: "裁剪空白",
emptyStringAsNull: "空字串作為 NULL", emptyStringAsNull: "空字串作為 NULL",
conflictPolicy: "主鍵衝突時",
conflictError: "停止並回報錯誤",
skipDuplicateRows: "略過重複列",
updateExistingRows: "更新既有列",
updateExistingRequiresExistingTable: "衝突更新僅適用於既有目標資料表。",
updateExistingUnsupportedDialect: "目前資料庫不支援衝突更新。",
updateExistingRequiresPrimaryKey: "衝突更新需要目標資料表的主鍵中繼資料。",
updateExistingRequiresMappedPrimaryKey: "衝突更新需要映射所有主鍵欄位,缺少:{columns}",
updateExistingRequiresMappedValue: "衝突更新需要至少映射一個非主鍵欄位。",
jsonShape: "JSON 列結構", jsonShape: "JSON 列結構",
jsonShapeAuto: "自動偵測", jsonShapeAuto: "自動偵測",
jsonShapeObjects: "物件", jsonShapeObjects: "物件",
+1
View File
@@ -1094,6 +1094,7 @@ export type {
TransferOwnershipPolicy, TransferOwnershipPolicy,
TransferOwnershipPreview, TransferOwnershipPreview,
TableImportMode, TableImportMode,
TableImportConflictPolicy,
TableImportStatus, TableImportStatus,
TableImportSourceFormat, TableImportSourceFormat,
TableImportJsonShape, TableImportJsonShape,
+2
View File
@@ -5375,6 +5375,7 @@ export async function sortTablesByFkDependency(options: SortTablesByFkOptions):
// --- Table File Import --- // --- Table File Import ---
export type TableImportMode = "append" | "truncate"; export type TableImportMode = "append" | "truncate";
export type TableImportConflictPolicy = "error" | "skip" | "updateExisting";
export type TableImportStatus = "running" | "done" | "error" | "cancelled"; export type TableImportStatus = "running" | "done" | "error" | "cancelled";
export type TableImportPhase = "preparing" | "detectingEncoding" | "reading" | "writing" | "finalizing" | "done"; export type TableImportPhase = "preparing" | "detectingEncoding" | "reading" | "writing" | "finalizing" | "done";
export type TableImportSourceFormat = "csv" | "tsv" | "delimited" | "json" | "excel" | "sql"; export type TableImportSourceFormat = "csv" | "tsv" | "delimited" | "json" | "excel" | "sql";
@@ -5451,6 +5452,7 @@ export interface TableImportRequest {
dateTimeFormat?: string; dateTimeFormat?: string;
preparedSource?: TableImportPreparedSource | null; preparedSource?: TableImportPreparedSource | null;
retainSource?: boolean; retainSource?: boolean;
conflictPolicy?: TableImportConflictPolicy;
skipDuplicateRows?: boolean; skipDuplicateRows?: boolean;
} }
@@ -406,6 +406,7 @@ fn import_request(
date_time_format: None, date_time_format: None,
prepared_source: None, prepared_source: None,
retain_source: false, retain_source: false,
conflict_policy: None,
skip_duplicate_rows: false, skip_duplicate_rows: false,
} }
} }
+466 -88
View File
@@ -29,7 +29,8 @@ use crate::sql::SqlParsingOptions;
use crate::sql_file_import::{SqlFileStreamDecoder, StreamingSqlFileSplitter}; use crate::sql_file_import::{SqlFileStreamDecoder, StreamingSqlFileSplitter};
use crate::transfer::{ use crate::transfer::{
escape_value_typed, execute_on_pool, generate_insert_typed_from_value_rows, get_columns_for_transfer, escape_value_typed, execute_on_pool, generate_insert_typed_from_value_rows, get_columns_for_transfer,
normalize_integer_literal, normalize_thousands_numeric_literal, qualified_table, quote_identifier, SqlBatchLimits, normalize_integer_literal, normalize_thousands_numeric_literal, qualified_table, quote_identifier,
supports_primary_key_upsert, SqlBatchLimits,
}; };
pub const DEFAULT_PREVIEW_LIMIT: usize = 50; pub const DEFAULT_PREVIEW_LIMIT: usize = 50;
@@ -245,6 +246,15 @@ pub struct TableImportPreviewRequest {
pub preview_limit: Option<usize>, pub preview_limit: Option<usize>,
} }
#[derive(Debug, Clone, Copy, Serialize, Deserialize, Default, PartialEq, Eq)]
#[serde(rename_all = "camelCase")]
pub enum TableImportConflictPolicy {
#[default]
Error,
Skip,
UpdateExisting,
}
#[derive(Debug, Clone, Serialize, Deserialize)] #[derive(Debug, Clone, Serialize, Deserialize)]
#[serde(rename_all = "camelCase")] #[serde(rename_all = "camelCase")]
pub struct TableImportRequest { pub struct TableImportRequest {
@@ -271,10 +281,83 @@ pub struct TableImportRequest {
pub prepared_source: Option<TableImportPreparedSource>, pub prepared_source: Option<TableImportPreparedSource>,
#[serde(default)] #[serde(default)]
pub retain_source: bool, pub retain_source: bool,
/// Explicit conflict behavior for new callers. When omitted, the legacy
/// `skip_duplicate_rows` flag remains authoritative for compatibility.
#[serde(default, skip_serializing_if = "Option::is_none")]
pub conflict_policy: Option<TableImportConflictPolicy>,
#[serde(default)] #[serde(default)]
pub skip_duplicate_rows: bool, pub skip_duplicate_rows: bool,
} }
impl TableImportRequest {
fn effective_conflict_policy(&self) -> TableImportConflictPolicy {
self.conflict_policy.unwrap_or(if self.skip_duplicate_rows {
TableImportConflictPolicy::Skip
} else {
TableImportConflictPolicy::Error
})
}
}
fn validate_update_existing_target(
conflict_policy: TableImportConflictPolicy,
db_type: &DatabaseType,
create_table: bool,
) -> Result<(), String> {
if conflict_policy != TableImportConflictPolicy::UpdateExisting {
return Ok(());
}
if create_table {
return Err("Update-existing import requires an existing target table".to_string());
}
if !supports_primary_key_upsert(db_type) {
return Err(format!("Update-existing import conflict policy is not supported for {}", db_type.as_str()));
}
Ok(())
}
fn update_existing_primary_key_columns(
conflict_policy: TableImportConflictPolicy,
mappings: &[TableImportColumnMapping],
target_columns: &[crate::db::ColumnInfo],
) -> Result<Vec<String>, String> {
if conflict_policy != TableImportConflictPolicy::UpdateExisting {
return Ok(Vec::new());
}
let primary_key_columns = target_columns.iter().filter(|column| column.is_primary_key).collect::<Vec<_>>();
if primary_key_columns.is_empty() {
return Err("Update-existing import requires target primary-key metadata".to_string());
}
let mut mapped_primary_keys = Vec::with_capacity(primary_key_columns.len());
let mut missing_primary_keys = Vec::new();
for primary_key in primary_key_columns {
if let Some(mapping) =
mappings.iter().find(|mapping| mapping.target_column.eq_ignore_ascii_case(&primary_key.name))
{
mapped_primary_keys.push(mapping.target_column.clone());
} else {
missing_primary_keys.push(primary_key.name.clone());
}
}
if !missing_primary_keys.is_empty() {
return Err(format!(
"Update-existing import requires every primary-key column to be mapped; missing: {}",
missing_primary_keys.join(", ")
));
}
let has_mapped_non_primary_key = mappings.iter().any(|mapping| {
!mapped_primary_keys.iter().any(|primary_key| primary_key.eq_ignore_ascii_case(&mapping.target_column))
});
if !has_mapped_non_primary_key {
return Err("Update-existing import requires at least one mapped non-primary-key column".to_string());
}
Ok(mapped_primary_keys)
}
#[derive(Debug, Clone, Serialize, Deserialize)] #[derive(Debug, Clone, Serialize, Deserialize)]
#[serde(rename_all = "camelCase")] #[serde(rename_all = "camelCase")]
pub struct TableImportPreparedSource { pub struct TableImportPreparedSource {
@@ -4347,7 +4430,8 @@ fn build_import_insert_batches_with_plan(
schema: &str, schema: &str,
db_type: &DatabaseType, db_type: &DatabaseType,
kingbase_oracle_mode: bool, kingbase_oracle_mode: bool,
skip_duplicate_rows: bool, conflict_policy: TableImportConflictPolicy,
primary_key_columns: &[String],
date_time_format: Option<&str>, date_time_format: Option<&str>,
hard_sql_bytes: Option<usize>, hard_sql_bytes: Option<usize>,
) -> Result<Vec<ImportSqlBatch>, String> { ) -> Result<Vec<ImportSqlBatch>, String> {
@@ -4355,16 +4439,33 @@ fn build_import_insert_batches_with_plan(
return Ok(Vec::new()); return Ok(Vec::new());
} }
let value_rows = import_value_rows_sql(rows, plan, db_type, kingbase_oracle_mode, date_time_format); let value_rows = import_value_rows_sql(rows, plan, db_type, kingbase_oracle_mode, date_time_format);
let batches = crate::data::transfer::generate_insert_typed_sql_batches_from_value_rows_with_options( let limits = SqlBatchLimits::for_database(db_type, rows.len()).with_hard_sql_bytes(hard_sql_bytes);
&plan.target_columns, let batches = match conflict_policy {
&value_rows, TableImportConflictPolicy::Error | TableImportConflictPolicy::Skip => {
table, crate::data::transfer::generate_insert_typed_sql_batches_from_value_rows_with_options(
schema, &plan.target_columns,
db_type, &value_rows,
None, table,
SqlBatchLimits::for_database(db_type, rows.len()).with_hard_sql_bytes(hard_sql_bytes), schema,
skip_duplicate_rows, db_type,
)?; None,
limits,
conflict_policy == TableImportConflictPolicy::Skip,
)?
}
TableImportConflictPolicy::UpdateExisting => {
crate::data::transfer::generate_primary_key_upsert_sql_batches_from_value_rows(
&plan.target_columns,
&value_rows,
table,
schema,
db_type,
None,
limits,
primary_key_columns,
)?
}
};
Ok(batches.into_iter().map(|(sql, row_count)| ImportSqlBatch { sql, row_count }).collect()) Ok(batches.into_iter().map(|(sql, row_count)| ImportSqlBatch { sql, row_count }).collect())
} }
@@ -4429,7 +4530,8 @@ fn build_import_execution_batches(
schema: &str, schema: &str,
db_type: &DatabaseType, db_type: &DatabaseType,
kingbase_oracle_mode: bool, kingbase_oracle_mode: bool,
skip_duplicate_rows: bool, conflict_policy: TableImportConflictPolicy,
primary_key_columns: &[String],
date_time_format: Option<&str>, date_time_format: Option<&str>,
hard_sql_bytes: Option<usize>, hard_sql_bytes: Option<usize>,
) -> Result<Vec<ImportSqlBatch>, String> { ) -> Result<Vec<ImportSqlBatch>, String> {
@@ -4441,12 +4543,13 @@ fn build_import_execution_batches(
schema, schema,
db_type, db_type,
kingbase_oracle_mode, kingbase_oracle_mode,
skip_duplicate_rows, conflict_policy,
primary_key_columns,
date_time_format, date_time_format,
hard_sql_bytes, hard_sql_bytes,
); );
} }
if *db_type == DatabaseType::CloudflareD1 { if *db_type == DatabaseType::CloudflareD1 && conflict_policy != TableImportConflictPolicy::UpdateExisting {
return crate::db::cloudflare_d1::build_import_insert_batches( return crate::db::cloudflare_d1::build_import_insert_batches(
rows, rows,
columns, columns,
@@ -4465,7 +4568,8 @@ fn build_import_execution_batches(
schema, schema,
db_type, db_type,
kingbase_oracle_mode, kingbase_oracle_mode,
skip_duplicate_rows, conflict_policy,
primary_key_columns,
date_time_format, date_time_format,
hard_sql_bytes, hard_sql_bytes,
) )
@@ -4664,7 +4768,8 @@ fn build_import_insert_batches_with_format(
schema, schema,
db_type, db_type,
kingbase_oracle_mode, kingbase_oracle_mode,
false, TableImportConflictPolicy::Error,
&[],
date_time_format, date_time_format,
None, None,
)?); )?);
@@ -5503,6 +5608,18 @@ fn import_batch_execution_policy(
} }
} }
fn completed_import_rows(
conflict_policy: TableImportConflictPolicy,
submitted_rows: usize,
affected_rows: u64,
) -> usize {
if conflict_policy == TableImportConflictPolicy::Skip {
affected_rows as usize
} else {
submitted_rows
}
}
#[allow(clippy::too_many_arguments)] #[allow(clippy::too_many_arguments)]
async fn execute_import_transaction( async fn execute_import_transaction(
state: &AppState, state: &AppState,
@@ -5639,14 +5756,15 @@ async fn execute_import_rows_batch(
postgres_copy_accumulator: &mut Option<PostgresCopyAccumulator>, postgres_copy_accumulator: &mut Option<PostgresCopyAccumulator>,
sqlite_append_transaction: &mut Option<SqliteAppendTransaction>, sqlite_append_transaction: &mut Option<SqliteAppendTransaction>,
kingbase_oracle_mode: bool, kingbase_oracle_mode: bool,
skip_duplicate_rows: bool, conflict_policy: TableImportConflictPolicy,
primary_key_columns: &[String],
date_time_format: Option<&str>, date_time_format: Option<&str>,
hard_sql_bytes: Option<usize>, hard_sql_bytes: Option<usize>,
db_write_ms: &mut u128, db_write_ms: &mut u128,
statement_count: &mut usize, statement_count: &mut usize,
) -> Result<usize, ImportRowsBatchError> { ) -> Result<usize, ImportRowsBatchError> {
let execution_policy = import_batch_execution_policy(mode, pending_truncate, db_type); let execution_policy = import_batch_execution_policy(mode, pending_truncate, db_type);
if !skip_duplicate_rows { if conflict_policy == TableImportConflictPolicy::Error {
if let Some((import_plan, bulk_plan)) = sqlserver_bulk_plans_for_rows(db_type, plan, sqlserver_bulk_plan, rows) if let Some((import_plan, bulk_plan)) = sqlserver_bulk_plans_for_rows(db_type, plan, sqlserver_bulk_plan, rows)
{ {
return execute_sqlserver_bulk_rows_batch( return execute_sqlserver_bulk_rows_batch(
@@ -5667,7 +5785,7 @@ async fn execute_import_rows_batch(
} }
// COPY is used only for plain scalar PostgreSQL rows and ordinary tables. Any unsupported // COPY is used only for plain scalar PostgreSQL rows and ordinary tables. Any unsupported
// value or table feature falls through to the portable INSERT generator below. // value or table feature falls through to the portable INSERT generator below.
if !skip_duplicate_rows if conflict_policy == TableImportConflictPolicy::Error
&& execution_policy.allow_postgres_copy && execution_policy.allow_postgres_copy
&& *db_type == DatabaseType::Postgres && *db_type == DatabaseType::Postgres
&& !rows && !rows
@@ -5711,7 +5829,8 @@ async fn execute_import_rows_batch(
schema, schema,
db_type, db_type,
kingbase_oracle_mode, kingbase_oracle_mode,
skip_duplicate_rows, conflict_policy,
primary_key_columns,
date_time_format, date_time_format,
hard_sql_bytes, hard_sql_bytes,
) )
@@ -5774,22 +5893,19 @@ async fn execute_import_rows_batch(
) )
.await .await
.map_err(|message| ImportRowsBatchError::with_rows_imported(rows_imported, message))?; .map_err(|message| ImportRowsBatchError::with_rows_imported(rows_imported, message))?;
return Ok(if skip_duplicate_rows { return Ok(rows_imported.saturating_add(completed_import_rows(
rows_imported.saturating_add(result.affected_rows as usize) conflict_policy,
} else { rows.len(),
rows_imported.saturating_add(rows.len()) result.affected_rows,
}); )));
} }
for batch in batches { for batch in batches {
ensure_import_write_allowed(import_id, is_cancelled, rows_imported).await?; ensure_import_write_allowed(import_id, is_cancelled, rows_imported).await?;
let result = execute_import_statement(state, pool_key, &batch.sql, db_write_ms, statement_count) let result = execute_import_statement(state, pool_key, &batch.sql, db_write_ms, statement_count)
.await .await
.map_err(|error| ImportRowsBatchError::with_rows_imported(rows_imported, error))?; .map_err(|error| ImportRowsBatchError::with_rows_imported(rows_imported, error))?;
rows_imported = if skip_duplicate_rows { rows_imported =
rows_imported.saturating_add(result.affected_rows as usize) rows_imported.saturating_add(completed_import_rows(conflict_policy, batch.row_count, result.affected_rows));
} else {
rows_imported.saturating_add(batch.row_count)
};
} }
Ok(rows_imported) Ok(rows_imported)
} }
@@ -6594,6 +6710,10 @@ where
let mut statement_count = 0usize; let mut statement_count = 0usize;
let batch_size = if request.batch_size == 0 { DEFAULT_BATCH_SIZE } else { request.batch_size }; let batch_size = if request.batch_size == 0 { DEFAULT_BATCH_SIZE } else { request.batch_size };
let kingbase_oracle_mode = kingbase_oracle_compatibility_mode(state, pool_key, db_type).await; let kingbase_oracle_mode = kingbase_oracle_compatibility_mode(state, pool_key, db_type).await;
let conflict_policy = request.effective_conflict_policy();
if let Err(error) = validate_update_existing_target(conflict_policy, db_type, request.create_table) {
return Err(emit_import_error(&mut progress_callback, request, 0, 0, started_at, error));
}
let source_format = match effective_source_format(&request.file_path, request.source_format) { let source_format = match effective_source_format(&request.file_path, request.source_format) {
Ok(format) => format, Ok(format) => format,
Err(error) => { Err(error) => {
@@ -6750,6 +6870,43 @@ where
create_table_sample = Some(parsed); create_table_sample = Some(parsed);
} }
let target_columns = match get_columns_for_transfer(
state,
pool_key,
&request.connection_id,
&request.database,
&request.schema,
&request.table,
None,
)
.await
{
Ok(columns) => columns,
Err(error) if conflict_policy == TableImportConflictPolicy::UpdateExisting => {
return Err(emit_import_error(
&mut progress_callback,
request,
0,
0,
started_at,
format!("Update-existing import could not load target primary-key metadata: {error}"),
));
}
Err(_) => Vec::new(),
};
let primary_key_columns =
match update_existing_primary_key_columns(conflict_policy, &request.mappings, &target_columns) {
Ok(columns) => columns,
Err(error) => {
return Err(emit_import_error(&mut progress_callback, request, 0, 0, started_at, error));
}
};
let mut target_column_types =
target_columns.iter().map(|column| (column.name.clone(), column.data_type.clone())).collect::<Vec<_>>();
if target_column_types.is_empty() {
target_column_types = created_column_types.clone().unwrap_or_default();
}
if source_format.is_delimited() { if source_format.is_delimited() {
let parsed = if let Some(parsed) = create_table_sample.clone().or_else(|| prepared_source.clone()) { let parsed = if let Some(parsed) = create_table_sample.clone().or_else(|| prepared_source.clone()) {
parsed parsed
@@ -6773,23 +6930,6 @@ where
let total_bytes = tokio::fs::metadata(&request.file_path).await.map(|metadata| metadata.len()).unwrap_or(0); let total_bytes = tokio::fs::metadata(&request.file_path).await.map(|metadata| metadata.len()).unwrap_or(0);
let mut target_column_types = get_columns_for_transfer(
state,
pool_key,
&request.connection_id,
&request.database,
&request.schema,
&request.table,
None,
)
.await
.unwrap_or_default()
.into_iter()
.map(|column| (column.name, column.data_type))
.collect::<Vec<_>>();
if target_column_types.is_empty() {
target_column_types = created_column_types.clone().unwrap_or_default();
}
let (resolved_encoding, _) = let (resolved_encoding, _) =
validated_text_encoding.ok_or_else(|| "Delimited import encoding was not validated".to_string())?; validated_text_encoding.ok_or_else(|| "Delimited import encoding was not validated".to_string())?;
let mut streaming_options = import_parse_options.clone(); let mut streaming_options = import_parse_options.clone();
@@ -6949,7 +7089,8 @@ where
&mut postgres_copy_accumulator, &mut postgres_copy_accumulator,
&mut sqlite_append_transaction, &mut sqlite_append_transaction,
kingbase_oracle_mode, kingbase_oracle_mode,
request.skip_duplicate_rows, conflict_policy,
&primary_key_columns,
request.date_time_format.as_deref(), request.date_time_format.as_deref(),
import_sql_hard_limit, import_sql_hard_limit,
&mut db_write_ms, &mut db_write_ms,
@@ -7141,23 +7282,6 @@ where
let effective_batch_size = effective_import_batch_size(db_type, batch_size); let effective_batch_size = effective_import_batch_size(db_type, batch_size);
let expected_columns = let expected_columns =
create_table_sample.as_ref().or(prepared_source.as_ref()).map(|source| source.columns.clone()); create_table_sample.as_ref().or(prepared_source.as_ref()).map(|source| source.columns.clone());
let mut target_column_types = get_columns_for_transfer(
state,
pool_key,
&request.connection_id,
&request.database,
&request.schema,
&request.table,
None,
)
.await
.unwrap_or_default()
.into_iter()
.map(|column| (column.name, column.data_type))
.collect::<Vec<_>>();
if target_column_types.is_empty() {
target_column_types = created_column_types.clone().unwrap_or_default();
}
let text_source_columns = textual_source_columns_for_import(&request.mappings, &target_column_types); let text_source_columns = textual_source_columns_for_import(&request.mappings, &target_column_types);
// No truncate, INSERT, or COPY may run until the selected worksheet parses to EOF. // No truncate, INSERT, or COPY may run until the selected worksheet parses to EOF.
let mut last_xlsx_read_bytes = 0u64; let mut last_xlsx_read_bytes = 0u64;
@@ -7452,7 +7576,8 @@ where
&mut postgres_copy_accumulator, &mut postgres_copy_accumulator,
&mut sqlite_append_transaction, &mut sqlite_append_transaction,
kingbase_oracle_mode, kingbase_oracle_mode,
request.skip_duplicate_rows, conflict_policy,
&primary_key_columns,
request.date_time_format.as_deref(), request.date_time_format.as_deref(),
import_sql_hard_limit, import_sql_hard_limit,
&mut db_write_ms, &mut db_write_ms,
@@ -7636,23 +7761,6 @@ where
started_at, started_at,
None, None,
)); ));
let mut target_column_types = get_columns_for_transfer(
state,
pool_key,
&request.connection_id,
&request.database,
&request.schema,
&request.table,
None,
)
.await
.unwrap_or_default()
.into_iter()
.map(|column| (column.name, column.data_type))
.collect::<Vec<_>>();
if target_column_types.is_empty() {
target_column_types = created_column_types.clone().unwrap_or_default();
}
let text_source_columns = textual_source_columns_for_import(&request.mappings, &target_column_types); let text_source_columns = textual_source_columns_for_import(&request.mappings, &target_column_types);
let effective_batch_size = effective_import_batch_size(db_type, batch_size); let effective_batch_size = effective_import_batch_size(db_type, batch_size);
let mut row_source = match ImportRowSource::open( let mut row_source = match ImportRowSource::open(
@@ -7801,7 +7909,8 @@ where
&mut postgres_copy_accumulator, &mut postgres_copy_accumulator,
&mut sqlite_append_transaction, &mut sqlite_append_transaction,
kingbase_oracle_mode, kingbase_oracle_mode,
request.skip_duplicate_rows, conflict_policy,
&primary_key_columns,
request.date_time_format.as_deref(), request.date_time_format.as_deref(),
import_sql_hard_limit, import_sql_hard_limit,
&mut db_write_ms, &mut db_write_ms,
@@ -7964,6 +8073,82 @@ mod tests {
assert_eq!(progress_json["totalRowsExact"], serde_json::json!(true)); assert_eq!(progress_json["totalRowsExact"], serde_json::json!(true));
} }
#[test]
fn table_import_conflict_policy_preserves_legacy_skip_requests() {
let request_json = serde_json::json!({
"importId": "import-1",
"connectionId": "connection-1",
"database": "db",
"schema": "public",
"table": "users",
"filePath": "users.csv",
"mappings": [],
"mode": "append",
"batchSize": 500
});
let default_request: TableImportRequest = serde_json::from_value(request_json.clone()).unwrap();
assert_eq!(default_request.effective_conflict_policy(), TableImportConflictPolicy::Error);
let mut legacy_skip_json = request_json.clone();
legacy_skip_json["skipDuplicateRows"] = serde_json::json!(true);
let legacy_skip: TableImportRequest = serde_json::from_value(legacy_skip_json).unwrap();
assert_eq!(legacy_skip.effective_conflict_policy(), TableImportConflictPolicy::Skip);
let mut explicit_json = request_json;
explicit_json["skipDuplicateRows"] = serde_json::json!(true);
explicit_json["conflictPolicy"] = serde_json::json!("updateExisting");
let explicit: TableImportRequest = serde_json::from_value(explicit_json).unwrap();
assert_eq!(explicit.effective_conflict_policy(), TableImportConflictPolicy::UpdateExisting);
}
#[test]
fn update_existing_progress_counts_source_rows_not_database_affected_rows() {
assert_eq!(completed_import_rows(TableImportConflictPolicy::Error, 3, 3), 3);
assert_eq!(completed_import_rows(TableImportConflictPolicy::Skip, 3, 1), 1);
// MySQL reports two affected rows for a changed row handled by ON DUPLICATE KEY UPDATE.
assert_eq!(completed_import_rows(TableImportConflictPolicy::UpdateExisting, 3, 6), 3);
}
#[test]
fn update_existing_requires_supported_dialect_and_complete_primary_key_mapping() {
let mappings = vec![
TableImportColumnMapping {
source_column: "tenant".to_string(),
target_column: "tenant_id".to_string(),
target_data_type: None,
},
TableImportColumnMapping {
source_column: "name".to_string(),
target_column: "name".to_string(),
target_data_type: None,
},
];
let target_columns = vec![
crate::db::ColumnInfo { name: "tenant_id".to_string(), is_primary_key: true, ..Default::default() },
crate::db::ColumnInfo { name: "id".to_string(), is_primary_key: true, ..Default::default() },
crate::db::ColumnInfo { name: "name".to_string(), ..Default::default() },
];
let missing_key =
update_existing_primary_key_columns(TableImportConflictPolicy::UpdateExisting, &mappings, &target_columns)
.unwrap_err();
assert!(missing_key.contains("id"));
let missing_metadata = update_existing_primary_key_columns(
TableImportConflictPolicy::UpdateExisting,
&mappings,
&[crate::db::ColumnInfo { name: "name".to_string(), ..Default::default() }],
)
.unwrap_err();
assert!(missing_metadata.contains("primary-key metadata"));
let unsupported =
validate_update_existing_target(TableImportConflictPolicy::UpdateExisting, &DatabaseType::Oracle, false)
.unwrap_err();
assert!(unsupported.contains("not supported for oracle"));
}
#[test] #[test]
fn compiled_import_plan_reuses_source_indexes_and_target_types() { fn compiled_import_plan_reuses_source_indexes_and_target_types() {
let columns = vec!["name".to_string(), "id".to_string()]; let columns = vec!["name".to_string(), "id".to_string()];
@@ -8047,6 +8232,7 @@ mod tests {
effective_encoding: Some(TableImportTextEncoding::Utf8), effective_encoding: Some(TableImportTextEncoding::Utf8),
}), }),
skip_duplicate_rows: false, skip_duplicate_rows: false,
conflict_policy: None,
retain_source: false, retain_source: false,
}; };
@@ -9761,6 +9947,7 @@ mod tests {
date_time_format: None, date_time_format: None,
prepared_source: None, prepared_source: None,
skip_duplicate_rows: false, skip_duplicate_rows: false,
conflict_policy: None,
retain_source: false, retain_source: false,
}; };
let started_at = Instant::now(); let started_at = Instant::now();
@@ -10172,6 +10359,7 @@ mod tests {
date_time_format: None, date_time_format: None,
prepared_source: None, prepared_source: None,
skip_duplicate_rows: false, skip_duplicate_rows: false,
conflict_policy: None,
retain_source: false, retain_source: false,
}; };
@@ -10262,6 +10450,7 @@ mod tests {
date_time_format: None, date_time_format: None,
prepared_source: None, prepared_source: None,
skip_duplicate_rows: false, skip_duplicate_rows: false,
conflict_policy: None,
retain_source: false, retain_source: false,
}; };
@@ -10343,6 +10532,7 @@ mod tests {
date_time_format: None, date_time_format: None,
prepared_source: None, prepared_source: None,
skip_duplicate_rows: false, skip_duplicate_rows: false,
conflict_policy: None,
retain_source: false, retain_source: false,
}; };
@@ -11426,6 +11616,99 @@ mod tests {
]); ]);
} }
#[test]
fn import_conflict_policy_keeps_default_and_skip_sql_unchanged() {
let plan = CompiledImportPlan {
mapped_source_indexes: vec![0, 1],
target_columns: vec!["id".to_string(), "name".to_string()],
column_types: vec![Some("integer".to_string()), Some("text".to_string())],
};
let rows = vec![vec![serde_json::json!(1), serde_json::json!("Ada")]];
let default_batches = build_import_insert_batches_with_plan(
&rows,
&plan,
"users",
"public",
&DatabaseType::Postgres,
false,
TableImportConflictPolicy::Error,
&[],
None,
None,
)
.unwrap();
assert_eq!(default_batches[0].sql, "INSERT INTO \"public\".\"users\" (\"id\", \"name\") VALUES\n(1, 'Ada')");
let skip_batches = build_import_insert_batches_with_plan(
&rows,
&plan,
"users",
"public",
&DatabaseType::Postgres,
false,
TableImportConflictPolicy::Skip,
&[],
None,
None,
)
.unwrap();
assert_eq!(
skip_batches[0].sql,
"INSERT INTO \"public\".\"users\" (\"id\", \"name\") VALUES\n(1, 'Ada')\nON CONFLICT DO NOTHING"
);
}
#[test]
fn update_existing_sql_uses_composite_primary_key_and_only_updates_non_keys() {
let plan = CompiledImportPlan {
mapped_source_indexes: vec![0, 1, 2, 3],
target_columns: vec!["tenant_id".to_string(), "id".to_string(), "name".to_string(), "status".to_string()],
column_types: vec![None, None, None, None],
};
let rows = vec![vec![
serde_json::json!(7),
serde_json::json!(42),
serde_json::json!("Ada"),
serde_json::json!("active"),
]];
let postgres = build_import_insert_batches_with_plan(
&rows,
&plan,
"users",
"public",
&DatabaseType::Postgres,
false,
TableImportConflictPolicy::UpdateExisting,
&["tenant_id".to_string(), "id".to_string()],
None,
None,
)
.unwrap();
assert!(postgres[0].sql.contains("ON CONFLICT (\"tenant_id\", \"id\") DO UPDATE SET"));
assert!(postgres[0].sql.contains("\"name\" = EXCLUDED.\"name\""));
assert!(postgres[0].sql.contains("\"status\" = EXCLUDED.\"status\""));
assert!(!postgres[0].sql.contains("\"id\" = EXCLUDED.\"id\""));
assert_eq!(postgres[0].row_count, 1);
let mysql = build_import_insert_batches_with_plan(
&rows,
&plan,
"users",
"",
&DatabaseType::Mysql,
false,
TableImportConflictPolicy::UpdateExisting,
&["tenant_id".to_string(), "id".to_string()],
None,
None,
)
.unwrap();
assert!(mysql[0].sql.contains("ON DUPLICATE KEY UPDATE `name` = VALUES(`name`), `status` = VALUES(`status`)"));
assert!(!mysql[0].sql.contains("`id` = VALUES(`id`)"));
}
#[test] #[test]
fn iris_import_uses_single_row_values_statements() { fn iris_import_uses_single_row_values_statements() {
let mappings = vec![TableImportColumnMapping { let mappings = vec![TableImportColumnMapping {
@@ -11698,7 +11981,8 @@ mod tests {
"dbo", "dbo",
&DatabaseType::SqlServer, &DatabaseType::SqlServer,
false, false,
false, TableImportConflictPolicy::Error,
&[],
None, None,
None, None,
) )
@@ -12024,7 +12308,8 @@ mod tests {
&mut sqlite_append_transaction, &mut sqlite_append_transaction,
&mut postgres_copy_accumulator, &mut postgres_copy_accumulator,
false, false,
false, TableImportConflictPolicy::Error,
&[],
None, None,
None, None,
&mut db_write_ms, &mut db_write_ms,
@@ -12242,7 +12527,8 @@ mod tests {
&mut self.postgres_copy_accumulator, &mut self.postgres_copy_accumulator,
&mut self.transaction, &mut self.transaction,
false, false,
false, TableImportConflictPolicy::Error,
&[],
None, None,
None, None,
&mut self.db_write_ms, &mut self.db_write_ms,
@@ -12368,6 +12654,7 @@ mod tests {
date_time_format: None, date_time_format: None,
prepared_source: None, prepared_source: None,
skip_duplicate_rows: false, skip_duplicate_rows: false,
conflict_policy: None,
retain_source: false, retain_source: false,
}; };
@@ -12395,6 +12682,97 @@ mod tests {
); );
} }
#[tokio::test]
async fn sqlite_update_existing_import_counts_updated_rows_once() {
let dir = tempfile::tempdir().unwrap();
let storage = crate::persistence::test_storage::open(&dir.path().join("storage.db")).await.unwrap();
let state = AppState::new(storage);
let connection_id = "sqlite-update-existing";
let pool_key = format!("{connection_id}:session:import");
let database_path = dir.path().join("target.db");
let sqlite = crate::db::sqlite::connect_path_create_if_missing(database_path.to_str().unwrap()).await.unwrap();
crate::db::sqlite::execute_query(
&sqlite,
"CREATE TABLE items (tenant_id INTEGER, id INTEGER, name TEXT, PRIMARY KEY (tenant_id, id))",
)
.await
.unwrap();
crate::db::sqlite::execute_query(&sqlite, "INSERT INTO items (tenant_id, id, name) VALUES (1, 1, 'Old')")
.await
.unwrap();
state
.update_connection_pools(|connections| {
connections.insert(pool_key.clone(), PoolKind::Sqlite(sqlite.clone()));
})
.await;
let config: ConnectionConfig = serde_json::from_value(serde_json::json!({
"id": connection_id,
"name": "SQLite update-existing test",
"db_type": "sqlite",
"host": "",
"port": 0,
"username": "",
"password": "",
"database": database_path.to_string_lossy()
}))
.unwrap();
state.configs.write().await.insert(connection_id.to_string(), config);
let data_path = dir.path().join("rows.csv");
std::fs::write(&data_path, b"tenant_id,id,name\n1,1,Updated\n1,2,Inserted\n").unwrap();
let request = TableImportRequest {
import_id: "sqlite-update-existing".to_string(),
connection_id: connection_id.to_string(),
database: String::new(),
schema: String::new(),
table: "items".to_string(),
file_path: data_path.to_string_lossy().to_string(),
source_ref: None,
source_format: Some(TableImportSourceFormat::Csv),
parse_options: TableImportParseOptions::default(),
mappings: ["tenant_id", "id", "name"]
.into_iter()
.map(|column| TableImportColumnMapping {
source_column: column.to_string(),
target_column: column.to_string(),
target_data_type: None,
})
.collect(),
mode: TableImportMode::Append,
create_table: false,
batch_size: 2,
date_time_format: None,
prepared_source: None,
retain_source: false,
conflict_policy: Some(TableImportConflictPolicy::UpdateExisting),
skip_duplicate_rows: false,
};
let summary = import_table_file_core(
&state,
&request,
&DatabaseType::Sqlite,
&pool_key,
|_| Box::pin(async { false }),
|_| {},
)
.await
.unwrap();
assert_eq!(summary.rows_imported, 2);
let rows =
crate::db::sqlite::execute_query(&sqlite, "SELECT tenant_id, id, name FROM items ORDER BY tenant_id, id")
.await
.unwrap()
.rows;
assert_eq!(
rows,
vec![
vec![serde_json::json!(1), serde_json::json!(1), serde_json::json!("Updated")],
vec![serde_json::json!(1), serde_json::json!(2), serde_json::json!("Inserted")],
]
);
}
#[tokio::test] #[tokio::test]
async fn preview_missing_source_fails_before_parsing() { async fn preview_missing_source_fails_before_parsing() {
let path = std::env::temp_dir().join(format!("dbx-missing-import-{}.csv", uuid::Uuid::new_v4())); let path = std::env::temp_dir().join(format!("dbx-missing-import-{}.csv", uuid::Uuid::new_v4()));
+119
View File
@@ -4036,6 +4036,76 @@ fn uses_mysql_style_upsert(db_type: &DatabaseType) -> bool {
matches!(db_type, DatabaseType::Mysql | DatabaseType::Doris | DatabaseType::StarRocks | DatabaseType::OpenGauss) matches!(db_type, DatabaseType::Mysql | DatabaseType::Doris | DatabaseType::StarRocks | DatabaseType::OpenGauss)
} }
pub(crate) fn supports_primary_key_upsert(db_type: &DatabaseType) -> bool {
matches!(
db_type,
DatabaseType::Postgres
| DatabaseType::Kingbase
| DatabaseType::Sqlite
| DatabaseType::CloudflareD1
| DatabaseType::DuckDb
| DatabaseType::Mysql
| DatabaseType::Doris
| DatabaseType::StarRocks
| DatabaseType::OpenGauss
)
}
fn primary_key_upsert_clause(
columns: &[String],
pk_columns: &[String],
db_type: &DatabaseType,
quote_target_column_names: bool,
) -> Result<String, String> {
if pk_columns.is_empty() {
return Err("Update-existing import requires target primary-key metadata".to_string());
}
if pk_columns.iter().any(|primary_key| !columns.iter().any(|column| column.eq_ignore_ascii_case(primary_key))) {
return Err("Update-existing import requires every target primary-key column to be mapped".to_string());
}
let non_pk_columns = columns
.iter()
.filter(|column| !pk_columns.iter().any(|primary_key| column.eq_ignore_ascii_case(primary_key)))
.collect::<Vec<_>>();
if non_pk_columns.is_empty() {
return Err("Update-existing import requires at least one mapped non-primary-key column".to_string());
}
if (is_postgres_transfer_dialect(db_type) && !matches!(db_type, DatabaseType::OpenGauss))
|| matches!(db_type, DatabaseType::Sqlite | DatabaseType::CloudflareD1 | DatabaseType::DuckDb)
{
let primary_keys = pk_columns
.iter()
.map(|column| transfer_column_identifier(column, db_type, quote_target_column_names))
.collect::<Vec<_>>()
.join(", ");
let updates = non_pk_columns
.iter()
.map(|column| {
let column = transfer_column_identifier(column, db_type, quote_target_column_names);
format!("{column} = EXCLUDED.{column}")
})
.collect::<Vec<_>>()
.join(", ");
return Ok(format!("\nON CONFLICT ({primary_keys}) DO UPDATE SET {updates}"));
}
if uses_mysql_style_upsert(db_type) {
let updates = non_pk_columns
.iter()
.map(|column| {
let column = transfer_column_identifier(column, db_type, quote_target_column_names);
format!("{column} = VALUES({column})")
})
.collect::<Vec<_>>()
.join(", ");
return Ok(format!("\nON DUPLICATE KEY UPDATE {updates}"));
}
Err(format!("Update-existing import conflict policy is not supported for {}", db_type.as_str()))
}
#[allow(clippy::too_many_arguments)] #[allow(clippy::too_many_arguments)]
fn generate_upsert_typed_for_transfer( fn generate_upsert_typed_for_transfer(
columns: &[String], columns: &[String],
@@ -4747,6 +4817,55 @@ pub(crate) fn generate_insert_typed_sql_batches_from_value_rows_with_options(
Ok(statements) Ok(statements)
} }
#[allow(clippy::too_many_arguments)]
pub(crate) fn generate_primary_key_upsert_sql_batches_from_value_rows(
columns: &[String],
value_rows: &[String],
table: &str,
schema: &str,
db_type: &DatabaseType,
catalog: Option<&str>,
limits: SqlBatchLimits,
pk_columns: &[String],
) -> Result<Vec<(String, usize)>, String> {
if !supports_primary_key_upsert(db_type) {
return Err(format!("Update-existing import conflict policy is not supported for {}", db_type.as_str()));
}
let clause = primary_key_upsert_clause(columns, pk_columns, db_type, true)?;
let clause_bytes = sql_text_bytes(&clause, db_type);
let adjusted_limits = SqlBatchLimits {
max_rows: limits.max_rows,
target_sql_bytes: limits.target_sql_bytes.saturating_sub(clause_bytes).max(1),
hard_sql_bytes: limits.hard_sql_bytes.map(|limit| limit.saturating_sub(clause_bytes).max(1)),
};
let batches = generate_insert_sql_batches_from_value_rows(
columns,
value_rows,
table,
schema,
db_type,
catalog,
adjusted_limits,
false,
true,
)?;
batches
.into_iter()
.map(|(mut sql, row_count)| {
sql.push_str(&clause);
if limits.hard_sql_bytes.is_some_and(|limit| sql_text_bytes(&sql, db_type) > limit) {
return Err(format!(
"SQL batch with update-existing conflict handling exceeds the {} byte hard limit",
limits.hard_sql_bytes.unwrap_or_default()
));
}
Ok((sql, row_count))
})
.collect()
}
#[allow(clippy::too_many_arguments)] #[allow(clippy::too_many_arguments)]
fn generate_insert_typed_sql_batches_for_transfer( fn generate_insert_typed_sql_batches_for_transfer(
columns: &[String], columns: &[String],
@@ -133,6 +133,7 @@ fn live_sqlserver_import_request(
date_time_format: None, date_time_format: None,
prepared_source: None, prepared_source: None,
retain_source: false, retain_source: false,
conflict_policy: None,
skip_duplicate_rows: false, skip_duplicate_rows: false,
} }
} }