diff --git a/apps/desktop/src/lib/__tests__/sql/sqlCompletion.functionFamilies.spec.ts b/apps/desktop/src/lib/__tests__/sql/sqlCompletion.functionFamilies.spec.ts new file mode 100644 index 0000000000..02fd008be9 --- /dev/null +++ b/apps/desktop/src/lib/__tests__/sql/sqlCompletion.functionFamilies.spec.ts @@ -0,0 +1,108 @@ +import { describe, expect, it } from "vitest"; +import { buildSqlCompletionItems, getSqlFunctionSignatureHelp } from "@/lib/sql/sqlCompletion"; +import type { DatabaseType } from "@/types/database"; + +/** 在 `SELECT xxx` 后触发补全,返回候选项;`prefix` 通常取函数名去掉最后一个字符 */ +function completionItems(prefix: string, databaseType: DatabaseType) { + const sql = `SELECT ${prefix}`; + return buildSqlCompletionItems(sql, sql.length, { databaseType, tables: [], columnsByTable: new Map() }); +} + +function functionItem(name: string, databaseType: DatabaseType) { + return completionItems(name.slice(0, -1), databaseType).find((item) => item.label === name); +} + +function functionLabels(databaseType: DatabaseType): string[] { + return completionItems("", databaseType) + .filter((item) => item.type === "function") + .map((item) => item.label); +} + +const ORACLE_FAMILY: DatabaseType[] = ["oracle", "oceanbase-oracle", "dameng", "yashandb", "oscar", "xugu"]; +const POSTGRES_FAMILY: DatabaseType[] = ["postgres", "redshift", "kingbase", "highgo", "uxdb", "vastbase", "gaussdb", "opengauss", "sundb"]; +const HIVE_SPARK_FAMILY: DatabaseType[] = ["hive", "spark", "databricks", "impala", "transwarp", "kyuubi", "argo"]; + +describe("SQL function completion across dialects", () => { + it.each(ORACLE_FAMILY)("offers Oracle built-in functions for %s", (databaseType) => { + expect(functionItem("NVL", databaseType)).toMatchObject({ type: "function", apply: expect.stringContaining("NVL(") }); + expect(functionItem("TO_CHAR", databaseType)).toMatchObject({ type: "function" }); + expect(functionItem("DECODE", databaseType)).toMatchObject({ type: "function" }); + expect(functionItem("ADD_MONTHS", databaseType)).toMatchObject({ type: "function" }); + expect(functionItem("LISTAGG", databaseType)).toMatchObject({ type: "function" }); + }); + + it.each(["oracle", "dameng"])("inserts Oracle syntax-sensitive functions through their templates for %s", (databaseType) => { + expect(functionItem("EXTRACT", databaseType)?.apply).toBe("EXTRACT(${unit} FROM ${datetime})"); + expect(functionItem("LISTAGG", databaseType)?.apply).toContain("WITHIN GROUP (ORDER BY ${order_by})"); + }); + + it("offers signature help for Oracle-only functions", () => { + const sql = "SELECT TO_CHAR("; + expect(getSqlFunctionSignatureHelp(sql, sql.length, "oracle")).not.toBeNull(); + expect(getSqlFunctionSignatureHelp(sql, sql.length, "mysql")).toBeNull(); + }); + + it.each(POSTGRES_FAMILY)("offers PostgreSQL family functions for %s", (databaseType) => { + expect(functionItem("DATE_TRUNC", databaseType)).toMatchObject({ type: "function" }); + expect(functionItem("SPLIT_PART", databaseType)).toMatchObject({ type: "function" }); + expect(functionItem("ARRAY_LENGTH", databaseType)).toMatchObject({ type: "function" }); + expect(functionItem("REGEXP_REPLACE", databaseType)).toMatchObject({ type: "function" }); + }); + + it("inserts PostgreSQL syntax-sensitive functions through their templates", () => { + expect(functionItem("POSITION", "postgres")?.apply).toBe("POSITION(${substring} IN ${string})"); + expect(functionItem("PERCENTILE_CONT", "kingbase")?.apply).toContain("WITHIN GROUP (ORDER BY ${expression})"); + }); + + it("offers DuckDB list and date functions on top of the PostgreSQL baseline", () => { + expect(functionItem("STRFTIME", "duckdb")).toMatchObject({ type: "function" }); + expect(functionItem("LIST_VALUE", "duckdb")).toMatchObject({ type: "function", apply: expect.stringContaining("LIST_VALUE(") }); + expect(functionItem("STRING_SPLIT", "duckdb")).toMatchObject({ type: "function" }); + // DuckDB 复用 PG 的语法模板 + expect(functionItem("POSITION", "duckdb")?.apply).toBe("POSITION(${substring} IN ${string})"); + // PG 基线函数同样可用 + expect(functionItem("DATE_TRUNC", "duckdb")).toMatchObject({ type: "function" }); + }); + + it.each(HIVE_SPARK_FAMILY)("offers Hive/Spark family functions for %s", (databaseType) => { + expect(functionItem("GET_JSON_OBJECT", databaseType)).toMatchObject({ type: "function" }); + expect(functionItem("COLLECT_LIST", databaseType)).toMatchObject({ type: "function" }); + expect(functionItem("EXPLODE", databaseType)).toMatchObject({ type: "function" }); + expect(functionItem("EXTRACT", databaseType)?.apply).toBe("EXTRACT(${field} FROM ${source})"); + }); + + it.each(["trino", "prestosql"])("offers Trino/Presto functions for %s", (databaseType) => { + expect(functionItem("DATE_DIFF", databaseType)).toMatchObject({ type: "function" }); + expect(functionItem("REGEXP_EXTRACT", databaseType)).toMatchObject({ type: "function" }); + expect(functionItem("ARRAY_JOIN", databaseType)).toMatchObject({ type: "function" }); + }); + + it("shares the MySQL function table with MySQL-compatible engines", () => { + expect(functionItem("DATE_FORMAT", "goldendb")).toMatchObject({ type: "function" }); + expect(functionItem("GROUP_CONCAT", "goldendb")).toMatchObject({ type: "function" }); + expect(functionItem("IFNULL", "goldendb")).toMatchObject({ type: "function" }); + }); + + it("keeps dialect-specific functions out of other dialects", () => { + const mysqlFunctions = functionLabels("mysql"); + expect(mysqlFunctions).not.toContain("NVL"); + expect(mysqlFunctions).not.toContain("TO_CHAR"); + expect(mysqlFunctions).not.toContain("DATE_TRUNC"); + expect(mysqlFunctions).not.toContain("GET_JSON_OBJECT"); + + const oracleFunctions = functionLabels("oracle"); + expect(oracleFunctions).not.toContain("IFNULL"); + expect(oracleFunctions).not.toContain("DATE_FORMAT"); + expect(oracleFunctions).not.toContain("GROUP_CONCAT"); + expect(oracleFunctions).not.toContain("STRFTIME"); + + const postgresFunctions = functionLabels("postgres"); + expect(postgresFunctions).not.toContain("NVL"); + expect(postgresFunctions).not.toContain("IFNULL"); + expect(postgresFunctions).not.toContain("GET_JSON_OBJECT"); + + const trinoFunctions = functionLabels("trino"); + expect(trinoFunctions).not.toContain("NVL"); + expect(trinoFunctions).not.toContain("GET_JSON_OBJECT"); + }); +}); diff --git a/apps/desktop/src/lib/sql/functionSignatures/duckdb.ts b/apps/desktop/src/lib/sql/functionSignatures/duckdb.ts new file mode 100644 index 0000000000..45093bd834 --- /dev/null +++ b/apps/desktop/src/lib/sql/functionSignatures/duckdb.ts @@ -0,0 +1,62 @@ +import { POSTGRES_FUNCTION_SIGNATURES } from "./postgres"; + +/** + * DuckDB 内置函数签名:DuckDB 有意兼容 PostgreSQL 语法,因此以 PG 表作为基线, + * 再叠加 DuckDB 特有的列表/结构体/日期函数(同名函数以本表为准)。 + */ +export const DUCKDB_FUNCTION_SIGNATURES = new Map([ + ...POSTGRES_FUNCTION_SIGNATURES, + // 日期时间 + ["STRFTIME", ["date", "format"]], + ["STRPTIME", ["text", "format"]], + ["DATE_DIFF", ["part", "startdate", "enddate"]], + ["DATEDIFF", ["part", "startdate", "enddate"]], + ["DATE_ADD", ["date", "interval"]], + ["DATE_SUB", ["date", "interval"]], + ["EPOCH", ["timestamp"]], + ["EPOCH_MS", ["milliseconds"]], + ["TODAY", []], + // 列表 / 结构体 / Map + ["LIST_VALUE", ["...values"]], + ["LIST_EXTRACT", ["list", "index"]], + ["LIST_TRANSFORM", ["list", "lambda"]], + ["LIST_FILTER", ["list", "lambda"]], + ["LIST_SORT", ["list"]], + ["LIST_REVERSE", ["list"]], + ["LIST_CONTAINS", ["list", "element"]], + ["LIST_POSITION", ["list", "element"]], + ["ARRAY_LENGTH", ["list"]], + ["LEN", ["string"]], + ["RANGE", ["start", "stop", "step"]], + ["STRUCT_PACK", ["...name_value_pairs"]], + ["MAP_KEYS", ["map"]], + ["MAP_VALUES", ["map"]], + ["MAP_EXTRACT", ["map", "key"]], + ["STRING_SPLIT", ["string", "separator"]], + ["STRING_SPLIT_REGEX", ["string", "regex"]], + // 字符串 + ["REGEXP_EXTRACT", ["string", "pattern", "group"]], + ["REGEXP_EXTRACT_ALL", ["string", "pattern", "group"]], + ["REGEXP_FULL_MATCH", ["string", "pattern", "options"]], + ["PREFIX", ["string", "search_string"]], + ["SUFFIX", ["string", "search_string"]], + ["CONTAINS", ["string", "search_string"]], + ["LEVENSHTEIN", ["string1", "string2"]], + ["JACCARD", ["string1", "string2"]], + // 聚合 + ["ARG_MAX", ["arg", "val"]], + ["ARG_MIN", ["arg", "val"]], + ["ARBITRARY", ["expr"]], + ["FIRST", ["expr"]], + ["LAST", ["expr"]], + ["HISTOGRAM", ["expr"]], + ["MODE", ["expr"]], + ["MEDIAN", ["expr"]], + ["QUANTILE_CONT", ["expr", "pos"]], + ["QUANTILE_DISC", ["expr", "pos"]], + ["APPROX_COUNT_DISTINCT", ["expr"]], + // 系统信息 + ["CURRENT_SETTING", ["setting"]], + ["CURRENT_DATABASE", []], + ["CURRENT_SCHEMA", []], +]); diff --git a/apps/desktop/src/lib/sql/functionSignatures/hiveSpark.ts b/apps/desktop/src/lib/sql/functionSignatures/hiveSpark.ts new file mode 100644 index 0000000000..ae9b31b0f8 --- /dev/null +++ b/apps/desktop/src/lib/sql/functionSignatures/hiveSpark.ts @@ -0,0 +1,140 @@ +/** + * Hive / Spark SQL 家族内置函数签名。 + * + * 说明: + * - 适用于 Hive、Spark、Databricks、Impala、Transwarp Inceptor、Kyuubi、Argo 等 Hive SQL 兼容引擎。 + * - 只收录高频 SQL 内置函数,参数名按 Hive/Spark 官方文档口径,仅作为补全提示与签名提示,不做语法校验。 + * - 窗口函数(ROW_NUMBER/RANK/LAG 等)由 `WINDOW_FUNCTIONS` 单独提供,此处不重复。 + */ +export const HIVE_SPARK_FUNCTION_SIGNATURES = new Map([ + // 空值 / 条件 + ["NVL", ["expr", "replacement"]], + ["NVL2", ["expr", "if_not_null", "if_null"]], + ["IF", ["condition", "true_value", "false_value"]], + ["ISNULL", ["expr"]], + ["ISNOTNULL", ["expr"]], + ["COALESCE", ["...values"]], + ["NULLIF", ["value1", "value2"]], + ["GREATEST", ["...values"]], + ["LEAST", ["...values"]], + ["ZEROIFNULL", ["expr"]], + ["NULLIFZERO", ["expr"]], + ["ASSERT_TRUE", ["condition"]], + // 字符串 + ["CONCAT", ["...values"]], + ["CONCAT_WS", ["separator", "...values"]], + ["SUBSTR", ["str", "pos", "len"]], + ["SUBSTRING", ["str", "pos", "len"]], + ["LENGTH", ["str"]], + ["UPPER", ["str"]], + ["LOWER", ["str"]], + ["INITCAP", ["str"]], + ["TRIM", ["str"]], + ["LTRIM", ["trim_str", "str"]], + ["RTRIM", ["trim_str", "str"]], + ["LPAD", ["str", "len", "pad"]], + ["RPAD", ["str", "len", "pad"]], + ["REPEAT", ["str", "n"]], + ["REVERSE", ["str"]], + ["SPLIT", ["str", "pattern"]], + ["INSTR", ["str", "substr"]], + ["LOCATE", ["substr", "str", "pos"]], + ["REGEXP_EXTRACT", ["str", "pattern", "idx"]], + ["REGEXP_EXTRACT_ALL", ["str", "pattern"]], + ["REGEXP_REPLACE", ["str", "pattern", "replacement"]], + ["ASCII", ["str"]], + ["CHR", ["num"]], + ["SOUNDEX", ["str"]], + ["TRANSLATE", ["input", "from", "to"]], + ["BASE64", ["bin"]], + ["UNBASE64", ["str"]], + ["HEX", ["expr"]], + ["UNHEX", ["expr"]], + ["MD5", ["expr"]], + ["SHA2", ["expr", "bit_length"]], + ["CRC32", ["expr"]], + ["HASH", ["...values"]], + // 日期时间(CURRENT_DATE / CURRENT_TIMESTAMP 是关键字项,这里不重复) + ["NOW", []], + ["UNIX_TIMESTAMP", []], + ["TO_UNIX_TIMESTAMP", ["date", "pattern"]], + ["FROM_UNIXTIME", ["unixtime", "format"]], + ["DATE_FORMAT", ["date", "format"]], + ["DATE_ADD", ["start_date", "days"]], + ["DATE_SUB", ["start_date", "days"]], + ["DATEDIFF", ["enddate", "startdate"]], + ["DATE_TRUNC", ["format", "timestamp"]], + ["TRUNC", ["date", "format"]], + ["MONTHS_BETWEEN", ["date1", "date2"]], + ["ADD_MONTHS", ["date", "months"]], + ["LAST_DAY", ["date"]], + ["NEXT_DAY", ["date", "day_of_week"]], + ["TO_DATE", ["date", "format"]], + ["YEAR", ["date"]], + ["QUARTER", ["date"]], + ["MONTH", ["date"]], + ["DAY", ["date"]], + ["HOUR", ["date"]], + ["MINUTE", ["date"]], + ["SECOND", ["date"]], + ["WEEKOFYEAR", ["date"]], + ["DAYOFWEEK", ["date"]], + ["DAYOFMONTH", ["date"]], + ["DAYOFYEAR", ["date"]], + ["FROM_UTC_TIMESTAMP", ["ts", "tz"]], + ["TO_UTC_TIMESTAMP", ["ts", "tz"]], + // 集合 / JSON + ["GET_JSON_OBJECT", ["json", "path"]], + ["FROM_JSON", ["json", "schema"]], + ["TO_JSON", ["obj"]], + ["JSON_TUPLE", ["json", "...paths"]], + ["EXPLODE", ["array"]], + ["POSEXPLODE", ["array"]], + ["INLINE", ["array_of_struct"]], + ["SIZE", ["array"]], + ["ARRAY", ["...values"]], + ["SORT_ARRAY", ["array", "ascending"]], + ["ARRAY_CONTAINS", ["array", "value"]], + ["MAP", ["...key_value_pairs"]], + ["MAP_KEYS", ["map"]], + ["MAP_VALUES", ["map"]], + ["STRUCT", ["...values"]], + ["NAMED_STRUCT", ["name", "...name_value_pairs"]], + // 聚合 / 分析 + ["COLLECT_LIST", ["expr"]], + ["COLLECT_SET", ["expr"]], + ["PERCENTILE", ["expr", "percentage"]], + ["PERCENTILE_APPROX", ["expr", "percentage"]], + ["APPROX_COUNT_DISTINCT", ["expr"]], + ["HISTOGRAM_NUMERIC", ["expr", "bins"]], + ["STDDEV", ["expr"]], + ["VARIANCE", ["expr"]], + ["CORR", ["expr1", "expr2"]], + ["COVAR_POP", ["expr1", "expr2"]], + ["COVAR_SAMP", ["expr1", "expr2"]], + // 数学 + ["ABS", ["expr"]], + ["CEIL", ["expr"]], + ["FLOOR", ["expr"]], + ["ROUND", ["expr", "d"]], + ["POWER", ["a", "b"]], + ["SQRT", ["expr"]], + ["EXP", ["expr"]], + ["LN", ["expr"]], + ["LOG", ["base", "expr"]], + ["LOG10", ["expr"]], + ["LOG2", ["expr"]], + ["SIGN", ["expr"]], + ["PMOD", ["a", "b"]], + ["FACTORIAL", ["n"]], + ["BIN", ["expr"]], + ["CONV", ["num", "from_base", "to_base"]], + // 高阶函数 + ["TRANSFORM", ["array", "func"]], + ["FILTER", ["array", "func"]], + ["EXISTS", ["array", "func"]], + ["AGGREGATE", ["array", "start", "merge", "finish"]], +]); + +/** Hive/Spark 中参数语法特殊的函数,补全时按模板插入 */ +export const HIVE_SPARK_FUNCTION_APPLY_TEMPLATES = new Map([["EXTRACT", "EXTRACT(${field} FROM ${source})"]]); diff --git a/apps/desktop/src/lib/sql/functionSignatures/oracle.ts b/apps/desktop/src/lib/sql/functionSignatures/oracle.ts new file mode 100644 index 0000000000..e2deec8d93 --- /dev/null +++ b/apps/desktop/src/lib/sql/functionSignatures/oracle.ts @@ -0,0 +1,147 @@ +/** + * Oracle 内置 SQL 函数签名(含 OceanBase Oracle 模式、达梦、崖山、神通、虚谷等 Oracle 兼容库)。 + * + * 说明: + * - 只收录「高频且语法稳定」的 SQL 内置函数,参数名按 Oracle 官方文档口径,仅作为补全提示与签名提示, + * 不做语法校验;PL/SQL 过程化内置包(DBMS_*、UTL_*)不在补全范围内。 + * - 与 `COMMON_SQL_FUNCTION_NAMES` 重复的函数(COUNT/SUM/SUBSTR/COALESCE 等)在本表中可以覆盖写法, + * 命中时以本表为准(`activeFunctionSignatures` 中方言表后合并)。 + * - 窗口函数(ROW_NUMBER/RANK/LAG 等)由 `WINDOW_FUNCTIONS` 单独提供,此处不重复。 + */ +export const ORACLE_FUNCTION_SIGNATURES = new Map([ + // 空值 / 条件 + ["NVL", ["expr1", "expr2"]], + ["NVL2", ["expr1", "expr2", "expr3"]], + ["DECODE", ["expr", "search", "result", "...default"]], + ["LNNVL", ["condition"]], + ["NANVL", ["value", "replacement"]], + ["GREATEST", ["...values"]], + ["LEAST", ["...values"]], + ["NULLIF", ["expr1", "expr2"]], + ["COALESCE", ["expr", "...exprs"]], + // 类型转换 + ["TO_CHAR", ["value", "format"]], + ["TO_DATE", ["char", "format"]], + ["TO_NUMBER", ["expr", "format"]], + ["TO_TIMESTAMP", ["char", "format"]], + ["TO_TIMESTAMP_TZ", ["char", "format"]], + ["TO_NCHAR", ["expr"]], + ["TO_CLOB", ["char"]], + ["TO_BLOB", ["raw"]], + ["TO_DSINTERVAL", ["char"]], + ["TO_YMINTERVAL", ["char"]], + ["TO_MULTI_BYTE", ["char"]], + ["TO_SINGLE_BYTE", ["char"]], + ["TREAT", ["expr AS type"]], + ["CAST", ["expr AS type"]], + ["RAWTOHEX", ["raw"]], + ["HEXTORAW", ["char"]], + ["SYS_GUID", []], + ["EMPTY_CLOB", []], + ["EMPTY_BLOB", []], + // 字符 + ["SUBSTR", ["string", "position", "length"]], + ["SUBSTRB", ["string", "position", "length"]], + ["INSTR", ["string", "substring", "position", "occurrence"]], + ["INSTRB", ["string", "substring", "position", "occurrence"]], + ["LENGTH", ["string"]], + ["LENGTHB", ["string"]], + ["LPAD", ["expr", "length", "pad"]], + ["RPAD", ["expr", "length", "pad"]], + ["TRIM", ["trim_source"]], + ["LTRIM", ["string", "set"]], + ["RTRIM", ["string", "set"]], + ["REPLACE", ["string", "search", "replacement"]], + ["TRANSLATE", ["expr", "from", "to"]], + ["UPPER", ["string"]], + ["LOWER", ["string"]], + ["INITCAP", ["string"]], + ["CONCAT", ["string1", "string2"]], + ["ASCII", ["char"]], + ["CHR", ["number"]], + ["SOUNDEX", ["char"]], + ["NLSSORT", ["char", "nlsparam"]], + ["UNISTR", ["string"]], + ["COMPOSE", ["string"]], + ["DECOMPOSE", ["string"]], + // 正则 + ["REGEXP_LIKE", ["source", "pattern", "match_param"]], + ["REGEXP_REPLACE", ["source", "pattern", "replacement", "position", "occurrence", "match_param"]], + ["REGEXP_SUBSTR", ["source", "pattern", "position", "occurrence", "match_param"]], + ["REGEXP_INSTR", ["source", "pattern", "position", "occurrence", "return_option", "match_param"]], + ["REGEXP_COUNT", ["source", "pattern", "position", "match_param"]], + // 数字 + ["ABS", ["number"]], + ["CEIL", ["number"]], + ["FLOOR", ["number"]], + ["ROUND", ["number", "decimals"]], + ["TRUNC", ["number", "decimals"]], + ["MOD", ["dividend", "divisor"]], + ["REMAINDER", ["n1", "n2"]], + ["POWER", ["base", "exponent"]], + ["SQRT", ["number"]], + ["EXP", ["number"]], + ["LN", ["number"]], + ["LOG", ["base", "number"]], + ["SIGN", ["number"]], + ["BITAND", ["expr1", "expr2"]], + ["WIDTH_BUCKET", ["expr", "min_value", "max_value", "num_buckets"]], + // 日期时间(SYSDATE / SYSTIMESTAMP / CURRENT_DATE / CURRENT_TIMESTAMP / LOCALTIMESTAMP / + // SESSIONTIMEZONE / DBTIMEZONE / USER / UID 由 `buildOracleSystemValueItems` 提供,插入时不带括号,这里不重复) + ["ADD_MONTHS", ["date", "months"]], + ["MONTHS_BETWEEN", ["date1", "date2"]], + ["LAST_DAY", ["date"]], + ["NEXT_DAY", ["date", "day"]], + ["NEW_TIME", ["date", "zone1", "zone2"]], + ["NUMTODSINTERVAL", ["n", "unit"]], + ["NUMTOYMINTERVAL", ["n", "unit"]], + ["SYS_EXTRACT_UTC", ["datetime"]], + ["TZ_OFFSET", ["time_zone"]], + ["FROM_TZ", ["timestamp", "time_zone"]], + // 语法特殊:参数不是普通逗号列表,apply 里用模板生成(见 ORACLE_FUNCTION_APPLY_TEMPLATES) + ["EXTRACT", ["unit FROM datetime"]], + ["POSITION", ["substring IN string"]], + // 聚合 / 分析(窗口函数见 WINDOW_FUNCTIONS) + ["LISTAGG", ["expr", "delimiter"]], + ["MEDIAN", ["expr"]], + ["STATS_MODE", ["expr"]], + ["RATIO_TO_REPORT", ["expr"]], + ["STDDEV", ["expr"]], + ["VARIANCE", ["expr"]], + ["CORR", ["expr1", "expr2"]], + ["COVAR_POP", ["expr1", "expr2"]], + ["COVAR_SAMP", ["expr1", "expr2"]], + ["GROUPING", ["expr"]], + ["GROUPING_ID", ["...exprs"]], + ["APPROX_COUNT_DISTINCT", ["expr"]], + // JSON(12c+) + ["JSON_VALUE", ["json", "path"]], + ["JSON_QUERY", ["json", "path"]], + ["JSON_EXISTS", ["json", "path"]], + ["JSON_OBJECT", ["key", "value", "...pairs"]], + ["JSON_ARRAY", ["...values"]], + ["JSON_MERGEPATCH", ["target", "patch"]], + ["JSON_EQUAL", ["expr1", "expr2"]], + // XML + ["XMLAGG", ["xml"]], + ["EXTRACTVALUE", ["xml", "xpath"]], + ["XMLELEMENT", ["name", "...content"]], + ["XMLFOREST", ["...values"]], + // 系统信息 / 杂项(USER / UID 由 Oracle 系统值项提供,此处不重复) + ["SYS_CONTEXT", ["namespace", "parameter"]], + ["USERENV", ["parameter"]], + ["ORA_HASH", ["expr", "max_bucket", "seed"]], + ["STANDARD_HASH", ["expr", "hash_method"]], + ["DUMP", ["expr"]], + ["VSIZE", ["expr"]], +]); + +/** + * Oracle 兼容族中语法与通用写法(逗号分隔参数)不同的函数, + * 这些函数在补全时会按模板插入,例如 EXTRACT 必须写成 `EXTRACT(YEAR FROM hiredate)`。 + */ +export const ORACLE_FUNCTION_APPLY_TEMPLATES = new Map([ + ["EXTRACT", "EXTRACT(${unit} FROM ${datetime})"], + ["POSITION", "POSITION(${substring} IN ${string})"], + ["LISTAGG", "LISTAGG(${expr}, ${delimiter}) WITHIN GROUP (ORDER BY ${order_by})"], +]); diff --git a/apps/desktop/src/lib/sql/functionSignatures/postgres.ts b/apps/desktop/src/lib/sql/functionSignatures/postgres.ts new file mode 100644 index 0000000000..d4d0f33b5c --- /dev/null +++ b/apps/desktop/src/lib/sql/functionSignatures/postgres.ts @@ -0,0 +1,153 @@ +/** + * PostgreSQL 家族内置函数签名。 + * + * 说明: + * - 这里同时作为 PG 兼容库(openGauss / GaussDB / Kingbase / Vastbase / HighGo / UXDB / SunDB / Redshift) + * 的共享函数表,`sqlCompletion.ts` 里按家族注册。 + * - 只收录高频 SQL 内置函数,参数名按 PostgreSQL 官方文档口径,仅作为补全提示与签名提示,不做语法校验。 + * - 窗口函数(ROW_NUMBER/RANK/LAG 等)由 `WINDOW_FUNCTIONS` 单独提供,此处不重复。 + */ +export const POSTGRES_FUNCTION_SIGNATURES = new Map([ + // 字符串 + ["LENGTH", ["string"]], + ["CHAR_LENGTH", ["string"]], + ["OCTET_LENGTH", ["string"]], + ["UPPER", ["string"]], + ["LOWER", ["string"]], + ["INITCAP", ["string"]], + ["CONCAT", ["...values"]], + ["CONCAT_WS", ["separator", "...values"]], + ["LEFT", ["string", "n"]], + ["RIGHT", ["string", "n"]], + ["LPAD", ["string", "length", "fill"]], + ["RPAD", ["string", "length", "fill"]], + ["TRIM", ["string"]], + ["BTRIM", ["string", "characters"]], + ["LTRIM", ["string"]], + ["RTRIM", ["string"]], + ["REPLACE", ["string", "from", "to"]], + ["REPEAT", ["string", "number"]], + ["SPLIT_PART", ["string", "delimiter", "n"]], + ["STRPOS", ["string", "substring"]], + ["POSITION", ["substring IN string"]], + ["SUBSTRING", ["string FROM start FOR count"]], + ["OVERLAY", ["string PLACING new FROM start FOR count"]], + ["FORMAT", ["format", "...values"]], + ["TRANSLATE", ["string", "from", "to"]], + ["ASCII", ["string"]], + ["CHR", ["code"]], + ["MD5", ["string"]], + ["ENCODE", ["data", "format"]], + ["DECODE", ["string", "format"]], + ["QUOTE_IDENT", ["string"]], + ["QUOTE_LITERAL", ["string"]], + // 正则 + ["REGEXP_REPLACE", ["string", "pattern", "replacement", "flags"]], + ["REGEXP_MATCHES", ["string", "pattern", "flags"]], + ["REGEXP_MATCH", ["string", "pattern", "flags"]], + ["REGEXP_SPLIT_TO_ARRAY", ["string", "pattern", "flags"]], + ["REGEXP_SPLIT_TO_TABLE", ["string", "pattern", "flags"]], + // 类型转换 + ["TO_CHAR", ["value", "format"]], + ["TO_DATE", ["text", "format"]], + ["TO_TIMESTAMP", ["text", "format"]], + ["TO_NUMBER", ["text", "format"]], + ["CAST", ["expression AS type"]], + // 日期时间(CURRENT_DATE / CURRENT_TIME / CURRENT_TIMESTAMP / LOCALTIME / LOCALTIMESTAMP 是关键字项,这里不重复) + ["NOW", []], + ["CLOCK_TIMESTAMP", []], + ["STATEMENT_TIMESTAMP", []], + ["TRANSACTION_TIMESTAMP", []], + ["DATE_TRUNC", ["field", "source"]], + ["DATE_PART", ["field", "source"]], + ["EXTRACT", ["field FROM source"]], + ["AGE", ["timestamp", "timestamp"]], + ["MAKE_DATE", ["year", "month", "day"]], + ["MAKE_TIME", ["hour", "min", "sec"]], + ["MAKE_TIMESTAMP", ["year", "month", "day", "hour", "min", "sec"]], + ["JUSTIFY_DAYS", ["interval"]], + ["JUSTIFY_HOURS", ["interval"]], + ["JUSTIFY_INTERVAL", ["interval"]], + ["PG_SLEEP", ["seconds"]], + // 数字 + ["ABS", ["number"]], + ["CEIL", ["number"]], + ["CEILING", ["number"]], + ["FLOOR", ["number"]], + ["ROUND", ["numeric", "int"]], + ["TRUNC", ["numeric", "int"]], + ["MOD", ["y", "x"]], + ["POWER", ["a", "b"]], + ["SQRT", ["number"]], + ["EXP", ["number"]], + ["LN", ["number"]], + ["LOG", ["b", "x"]], + ["LOG10", ["number"]], + ["SIGN", ["number"]], + ["RANDOM", []], + ["PI", []], + ["DEGREES", ["radians"]], + ["RADIANS", ["degrees"]], + ["WIDTH_BUCKET", ["operand", "low", "high", "count"]], + ["GCD", ["a", "b"]], + ["LCM", ["a", "b"]], + // 聚合 + ["ARRAY_AGG", ["expression"]], + ["STRING_AGG", ["expression", "delimiter"]], + ["JSON_AGG", ["expression"]], + ["JSONB_AGG", ["expression"]], + ["JSONB_OBJECT_AGG", ["key", "value"]], + ["BOOL_AND", ["expression"]], + ["BOOL_OR", ["expression"]], + ["EVERY", ["expression"]], + ["PERCENTILE_CONT", ["fraction"]], + ["PERCENTILE_DISC", ["fraction"]], + ["STDDEV", ["expression"]], + ["VARIANCE", ["expression"]], + ["CORR", ["y", "x"]], + ["COVAR_POP", ["y", "x"]], + ["COVAR_SAMP", ["y", "x"]], + // 数组 / JSON + ["UNNEST", ["array"]], + ["ARRAY_LENGTH", ["array", "dimension"]], + ["CARDINALITY", ["array"]], + ["ARRAY_POSITION", ["array", "element"]], + ["ARRAY_APPEND", ["array", "element"]], + ["ARRAY_REMOVE", ["array", "element"]], + ["ARRAY_CAT", ["array1", "array2"]], + ["ARRAY_TO_STRING", ["array", "delimiter"]], + ["STRING_TO_ARRAY", ["string", "delimiter"]], + ["GENERATE_SERIES", ["start", "stop", "step"]], + ["TO_JSON", ["value"]], + ["TO_JSONB", ["value"]], + ["JSONB_BUILD_OBJECT", ["key", "value", "...pairs"]], + ["JSON_BUILD_OBJECT", ["...key_value_pairs"]], + ["JSONB_SET", ["target", "path", "new_value"]], + ["JSONB_EXTRACT_PATH_TEXT", ["jsonb", "...path_elements"]], + ["JSONB_ARRAY_ELEMENTS", ["jsonb"]], + ["JSONB_ARRAY_ELEMENTS_TEXT", ["jsonb"]], + ["JSONB_EACH", ["jsonb"]], + ["JSONB_OBJECT_KEYS", ["jsonb"]], + // 系统信息 + ["GEN_RANDOM_UUID", []], + ["CURRENT_USER", []], + ["SESSION_USER", []], + ["CURRENT_SCHEMA", []], + ["CURRENT_DATABASE", []], + ["PG_BACKEND_PID", []], + ["PG_TYPEOF", ["any"]], +]); + +/** + * PostgreSQL 家族中语法与「逗号分隔参数」不同的函数,补全时按模板插入, + * 例如 `EXTRACT(YEAR FROM created_at)`、`PERCENTILE_CONT(0.5) WITHIN GROUP (ORDER BY salary)`。 + * DuckDB 复用 PG 风格语法,也使用这组模板。 + */ +export const POSTGRES_FUNCTION_APPLY_TEMPLATES = new Map([ + ["EXTRACT", "EXTRACT(${field} FROM ${source})"], + ["POSITION", "POSITION(${substring} IN ${string})"], + ["SUBSTRING", "SUBSTRING(${string} FROM ${start} FOR ${count})"], + ["OVERLAY", "OVERLAY(${string} PLACING ${new} FROM ${start} FOR ${count})"], + ["PERCENTILE_CONT", "PERCENTILE_CONT(${fraction}) WITHIN GROUP (ORDER BY ${expression})"], + ["PERCENTILE_DISC", "PERCENTILE_DISC(${fraction}) WITHIN GROUP (ORDER BY ${expression})"], +]); diff --git a/apps/desktop/src/lib/sql/functionSignatures/trino.ts b/apps/desktop/src/lib/sql/functionSignatures/trino.ts new file mode 100644 index 0000000000..6c44924045 --- /dev/null +++ b/apps/desktop/src/lib/sql/functionSignatures/trino.ts @@ -0,0 +1,123 @@ +/** + * Trino / PrestoSQL 内置函数签名。 + * + * 说明:只收录高频 SQL 内置函数,参数名按 Trino 官方文档口径,仅作为补全提示与签名提示,不做语法校验。 + * 窗口函数(ROW_NUMBER/RANK/LAG 等)由 `WINDOW_FUNCTIONS` 单独提供,此处不重复。 + */ +export const TRINO_FUNCTION_SIGNATURES = new Map([ + // 字符串 + ["CHR", ["n"]], + ["CODEPOINT", ["string"]], + ["CONCAT", ["...strings"]], + ["LENGTH", ["string"]], + ["UPPER", ["string"]], + ["LOWER", ["string"]], + ["LPAD", ["string", "size", "padstring"]], + ["RPAD", ["string", "size", "padstring"]], + ["LTRIM", ["string"]], + ["RTRIM", ["string"]], + ["TRIM", ["string"]], + ["REPLACE", ["string", "search", "replacement"]], + ["REVERSE", ["string"]], + ["SPLIT", ["string", "delimiter"]], + ["SPLIT_PART", ["string", "delimiter", "index"]], + ["STRPOS", ["string", "substring"]], + ["POSITION", ["substring IN string"]], + ["SUBSTR", ["string", "start", "length"]], + ["SUBSTRING", ["string FROM start FOR length"]], + ["LEVENSHTEIN_DISTANCE", ["string1", "string2"]], + ["HAMMING_DISTANCE", ["string1", "string2"]], + ["NORMALIZE", ["string", "form"]], + ["FORMAT", ["format", "...args"]], + ["REGEXP_EXTRACT", ["string", "pattern", "group"]], + ["REGEXP_EXTRACT_ALL", ["string", "pattern"]], + ["REGEXP_LIKE", ["string", "pattern"]], + ["REGEXP_REPLACE", ["string", "pattern", "replacement"]], + ["REGEXP_SPLIT", ["string", "pattern"]], + // 日期时间(CURRENT_* / LOCALTIME / LOCALTIMESTAMP 是关键字项,这里不重复) + ["NOW", []], + ["DATE", ["x"]], + ["DATE_ADD", ["unit", "value", "timestamp"]], + ["DATE_DIFF", ["unit", "timestamp1", "timestamp2"]], + ["DATE_FORMAT", ["timestamp", "format"]], + ["DATE_PARSE", ["string", "format"]], + ["DATE_TRUNC", ["unit", "x"]], + ["FORMAT_DATETIME", ["timestamp", "format"]], + ["PARSE_DATETIME", ["string", "format"]], + ["FROM_UNIXTIME", ["unixtime"]], + ["TO_UNIXTIME", ["timestamp"]], + ["TO_ISO8601", ["timestamp"]], + ["AT_TIMEZONE", ["timestamp", "zone"]], + ["WITH_TIMEZONE", ["timestamp", "zone"]], + ["EXTRACT", ["field FROM source"]], + // 数学 + ["ABS", ["x"]], + ["CEIL", ["x"]], + ["CEILING", ["x"]], + ["FLOOR", ["x"]], + ["ROUND", ["x", "d"]], + ["TRUNCATE", ["x", "d"]], + ["MOD", ["n", "m"]], + ["POWER", ["x", "p"]], + ["SQRT", ["x"]], + ["EXP", ["x"]], + ["LN", ["x"]], + ["LOG", ["b", "x"]], + ["LOG2", ["x"]], + ["LOG10", ["x"]], + ["SIGN", ["x"]], + ["RANDOM", ["n"]], + ["PI", []], + ["E", []], + ["DEGREES", ["x"]], + ["RADIANS", ["x"]], + // 数组 / Map + ["ARRAY_AGG", ["x"]], + ["ARRAY_DISTINCT", ["array"]], + ["ARRAY_JOIN", ["array", "delimiter"]], + ["CARDINALITY", ["array"]], + ["CONTAINS", ["array", "element"]], + ["ELEMENT_AT", ["array", "index"]], + ["FLATTEN", ["array"]], + ["FILTER", ["array", "lambda"]], + ["TRANSFORM", ["array", "lambda"]], + ["REDUCE", ["array", "initial", "lambda", "lambda"]], + ["SLICE", ["array", "start", "length"]], + ["SEQUENCE", ["start", "stop", "step"]], + ["ZIP", ["array1", "array2"]], + ["ZIP_WITH", ["array1", "array2", "lambda"]], + ["MAP_KEYS", ["map"]], + ["MAP_VALUES", ["map"]], + ["MAP_FROM_ENTRIES", ["entries"]], + // 聚合 + ["APPROX_DISTINCT", ["x", "e"]], + ["HISTOGRAM", ["x"]], + ["MAX_BY", ["x", "y"]], + ["MIN_BY", ["x", "y"]], + ["STDDEV", ["x"]], + ["VARIANCE", ["x"]], + ["CORR", ["y", "x"]], + // JSON / 转换 + ["JSON_EXTRACT", ["json", "path"]], + ["JSON_EXTRACT_SCALAR", ["json", "path"]], + ["JSON_FORMAT", ["json"]], + ["JSON_PARSE", ["string"]], + ["JSON_QUERY", ["json", "path"]], + ["JSON_VALUE", ["json", "path"]], + ["JSON_ARRAY_GET", ["json", "index"]], + ["JSON_SIZE", ["json", "path"]], + ["TRY", ["expression"]], + ["TRY_CAST", ["expression AS type"]], + ["IF", ["condition", "true_value", "false_value"]], + ["COALESCE", ["...values"]], + ["NULLIF", ["value1", "value2"]], + ["GREATEST", ["...values"]], + ["LEAST", ["...values"]], +]); + +/** Trino/Presto 中参数语法特殊的函数,补全时按模板插入 */ +export const TRINO_FUNCTION_APPLY_TEMPLATES = new Map([ + ["EXTRACT", "EXTRACT(${field} FROM ${source})"], + ["POSITION", "POSITION(${substring} IN ${string})"], + ["SUBSTRING", "SUBSTRING(${string} FROM ${start} FOR ${length})"], +]); diff --git a/apps/desktop/src/lib/sql/sqlCompletion.ts b/apps/desktop/src/lib/sql/sqlCompletion.ts index 881be5e638..5b2df1c35e 100644 --- a/apps/desktop/src/lib/sql/sqlCompletion.ts +++ b/apps/desktop/src/lib/sql/sqlCompletion.ts @@ -19,6 +19,11 @@ import { containsHan, orderedSubsequenceSpan, pinyinFirstLetters } from "@/lib/c import { quoteTableIdentifier, quoteTableIdentifierIfNeeded } from "@/lib/table/tableSelectSql"; import { driverProfileCompletionObjects, driverProfileCompletionTableMetadata, driverProfileCompletionTables, driverProfileRoutineSignatures } from "@/lib/database/driverProfileExtensions"; import { DORIS_FUNCTION_DOCS, DORIS_FUNCTION_SIGNATURES } from "@/lib/sql/doris/functions"; +import { DUCKDB_FUNCTION_SIGNATURES } from "@/lib/sql/functionSignatures/duckdb"; +import { HIVE_SPARK_FUNCTION_APPLY_TEMPLATES, HIVE_SPARK_FUNCTION_SIGNATURES } from "@/lib/sql/functionSignatures/hiveSpark"; +import { ORACLE_FUNCTION_APPLY_TEMPLATES, ORACLE_FUNCTION_SIGNATURES } from "@/lib/sql/functionSignatures/oracle"; +import { POSTGRES_FUNCTION_APPLY_TEMPLATES, POSTGRES_FUNCTION_SIGNATURES } from "@/lib/sql/functionSignatures/postgres"; +import { TRINO_FUNCTION_APPLY_TEMPLATES, TRINO_FUNCTION_SIGNATURES } from "@/lib/sql/functionSignatures/trino"; import { rejectsAliasReferenceInHaving } from "@/lib/database/databaseFeatureSupport"; import { isOracleCompletionDatabase } from "@/lib/sql/oracleCompletionSession"; import { normalizeSqlTableCompletionSchemaQualification, type SqlTableCompletionSchemaQualification } from "@/lib/sql/sqlCompletionSchemaQualification"; @@ -1030,17 +1035,6 @@ const SQL_FUNCTION_SIGNATURES = new Map([ ["JSON_OBJECTAGG", ["key", "value"]], ]); -const POSTGRES_FUNCTION_SIGNATURES = new Map([ - ["JSONB_BUILD_OBJECT", ["key", "value", "...pairs"]], - ["JSONB_AGG", ["expression"]], - ["TO_JSONB", ["value"]], - ["JSONB_SET", ["target", "path", "new_value"]], - ["ARRAY_AGG", ["expression"]], - ["STRING_AGG", ["expression", "delimiter"]], - ["GEN_RANDOM_UUID", []], - ["NOW", []], -]); - const MYSQL_FUNCTION_SIGNATURES = new Map([ ["CONVERT", ["expression", "type"]], ["DATE_FORMAT", ["date", "format"]], @@ -1160,7 +1154,7 @@ const MYSQL_FUNCTION_SIGNATURES = new Map([ ]); /** Keywords that may also exist as built-in functions; keep both completion entries. */ -const DUAL_ROLE_SQL_KEYWORDS = new Set(["LEFT", "RIGHT", "IF", "TRUNCATE", "REPEAT", "DATABASE", "SCHEMA", "USER", "CURRENT_USER"]); +const DUAL_ROLE_SQL_KEYWORDS = new Set(["LEFT", "RIGHT", "IF", "EXISTS", "TRUNCATE", "REPEAT", "DATABASE", "SCHEMA", "USER", "CURRENT_USER"]); const SQLITE_FUNCTION_SIGNATURES = new Map([ ["JSON_EXTRACT", ["json", "path"]], @@ -1232,9 +1226,14 @@ const MANTICORESEARCH_FUNCTION_SIGNATURES = new Map([ ["SECOND", ["timestamp"]], ]); +// 同一 SQL 方言族的兼容库共享一张函数表(以及函数插入模板),避免为每个兼容库重复维护同一份数据 +const MYSQL_COMPAT_FUNCTION_DATABASES = new Set(["mysql", "goldendb"]); +const POSTGRES_COMPAT_FUNCTION_DATABASES = new Set(["postgres", "redshift", "kingbase", "highgo", "uxdb", "vastbase", "gaussdb", "opengauss", "sundb"]); +const ORACLE_COMPAT_FUNCTION_DATABASES = new Set(["oracle", "oceanbase-oracle", "dameng", "yashandb", "oscar", "xugu"]); +const HIVE_SPARK_COMPAT_FUNCTION_DATABASES = new Set(["hive", "spark", "databricks", "impala", "transwarp", "kyuubi", "argo"]); +const TRINO_COMPAT_FUNCTION_DATABASES = new Set(["trino", "prestosql"]); + const DATABASE_FUNCTION_SIGNATURES: Partial>> = { - mysql: MYSQL_FUNCTION_SIGNATURES, - postgres: POSTGRES_FUNCTION_SIGNATURES, sqlite: SQLITE_FUNCTION_SIGNATURES, rqlite: SQLITE_FUNCTION_SIGNATURES, turso: SQLITE_FUNCTION_SIGNATURES, @@ -1243,8 +1242,34 @@ const DATABASE_FUNCTION_SIGNATURES: Partial, Map]>) { + for (const databaseType of databaseTypes) DATABASE_FUNCTION_SIGNATURES[databaseType] = signatures; +} + +/** + * 取当前方言的「函数插入模板」:EXTRACT / POSITION / LISTAGG / PERCENTILE_CONT 这类函数的参数 + * 不是逗号列表,必须按模板插入;模板与签名表同源于方言族。DuckDB 沿用 PG 风格语法,共用 PG 模板。 + */ +function functionApplyTemplate(databaseType: DatabaseType | undefined, name: string): string | undefined { + if (!databaseType) return undefined; + if (ORACLE_COMPAT_FUNCTION_DATABASES.has(databaseType)) return ORACLE_FUNCTION_APPLY_TEMPLATES.get(name); + if (databaseType === "duckdb" || POSTGRES_COMPAT_FUNCTION_DATABASES.has(databaseType)) return POSTGRES_FUNCTION_APPLY_TEMPLATES.get(name); + if (HIVE_SPARK_COMPAT_FUNCTION_DATABASES.has(databaseType)) return HIVE_SPARK_FUNCTION_APPLY_TEMPLATES.get(name); + if (TRINO_COMPAT_FUNCTION_DATABASES.has(databaseType)) return TRINO_FUNCTION_APPLY_TEMPLATES.get(name); + if (MYSQL_COMPAT_FUNCTION_DATABASES.has(databaseType)) return MYSQL_FUNCTION_APPLY_TEMPLATES.get(name); + return undefined; +} + const MYSQL_FUNCTION_APPLY_TEMPLATES = new Map([ ["DATE_ADD", "DATE_ADD(${date}, INTERVAL ${expr} ${unit})"], ["DATE_SUB", "DATE_SUB(${date}, INTERVAL ${expr} ${unit})"], @@ -5534,12 +5559,12 @@ function buildFunctionSnippetItems(prefix: string, functionDescriptions: Map 0 ? parameters.map((p) => `\${${applyGeneratedSqlTemplateKeywordCase(p, keywordCase)}}`).join(", ") : ""; - const mysqlApply = databaseType === "mysql" ? MYSQL_FUNCTION_APPLY_TEMPLATES.get(name) : undefined; + const applyTemplate = functionApplyTemplate(databaseType, name); items.push({ label: functionName, type: "function" as const, detail: activeFunctionDescriptions(databaseType).get(name) ?? functionDescriptions.get(name) ?? "function", - apply: mysqlApply ? `${functionName}${applyGeneratedSqlTemplateKeywordCase(mysqlApply.slice(name.length), keywordCase)}` : `${functionName}(${paramStr})`, + apply: applyTemplate ? `${functionName}${applyGeneratedSqlTemplateKeywordCase(applyTemplate.slice(name.length), keywordCase)}` : `${functionName}(${paramStr})`, boost: computeBoost(name, prefix) + 300, }); }