zipg
|
d715b16ba5
|
fix(sqlserver): wrap identity inserts with SET IDENTITY_INSERT
|
2026-09-20 23:49:28 +08:00 |
|
zipg
|
42980e5163
|
feat(import): stream SQL script parsing and drop the 100 MB cap
|
2026-09-20 23:47:29 +08:00 |
|
efeire
|
33afb0b432
|
feat(oceanbase-oracle): split execution time from client wait
|
2026-09-20 12:31:47 +08:00 |
|
efeire
|
a924285136
|
feat(transactions): add manual transactions to more execution paths
|
2026-09-20 11:07:11 +08:00 |
|
Loc Nguyen
|
a67bbb4ed5
|
feat(mongodb): support database dump and restore over legacy agent
|
2026-09-20 11:06:07 +08:00 |
|
zipg
|
e840da30fc
|
fix(transfer): use target column case for existing tables
|
2026-09-20 01:48:32 +08:00 |
|
 zipgandzipg
|
0adef6e16a
|
fix(export): 分页导出按列名对齐每一页,修复字段错位 (#9583)
Schemaless 引擎(Elasticsearch/Easysearch)的列集合是按每个响应里的文档现算的,
因此 select * 的不同分页可能给出不同的列顺序或列集合。查询结果导出把表头冻结在
第一页,却按位置写入后续页的行,导致第二页起的值落到错误的列里(#9522 反馈的
"导出字段错位",例如 beta 的值被写进 alpha 列)。
- 新增 export_column_remap/realign_page_cells/realign_page_rows:后续页按列名映射到
冻结表头,列名缺失时填 NULL,列顺序不同也能正确还原;
- 列类型同样按列名对齐,保证按页的日期/时间格式化使用正确的类型;
- spatial_values 随行一起对齐,保持与单元格的 SRID 对应关系;
- 第一页仍是表头来源,新增列(表头里没有的列)与被删除的列都会被忽略,行为与写出
表头保持一致。
Co-authored-by: zipg <4047349+zipg@users.noreply.github.com>
|
2026-09-19 15:23:22 +08:00 |
|
КГБ Lieu
|
30fe61d440
|
fix(sql-file): extend relational constraint bypass to Postgres and SQL Server
|
2026-09-19 09:47:58 +08:00 |
|
Elias
|
0db9756a57
|
fix(xugu): restore table import support
|
2026-09-19 09:24:16 +08:00 |
|
jearhe
|
f033e50ef5
|
fix(compare): preserve duplicate keys while removing sample overlap
|
2026-09-18 21:17:38 +08:00 |
|
t8y2
|
bc00431834
|
feat(core): split foundational capabilities into dedicated crates
|
2026-09-18 13:00:56 +08:00 |
|