flynn

> Fe use `checkstyle` check the code. This Pr is not compatible with it. > > You can see [this](https://doris.apache.org/zh-CN/community/developer-guide/java-format-code). Sorry, I don't descriple it clearlly, its some editor will...

+1 for C++ SDK

> Support for Iceberg Data Catalog `Catalog` is a good concept, but if want to introduce `catalog` into ClickHouse, we need refactor current metadata's structure, from `Database -> Table` to...

Question about these two table engines: currently implementation will get all data files when creating the table, but the iceberg/hudi table may have data written after table creation, while we...

> Please describe your change done.

I have met the same problem as well, but maybe doesn't because the file'name is non-English

Hi, I wonder which way will you use to support Hudi table? Because as I know, there is no Hudi C++ API.

@alexey-milovidov But parquet does not have `Date` and `DateTime` types: https://parquet.apache.org/docs/file-format/types/

For static sealed data, are you keep it as the same format that DiskANN needed? ``` The input data over which to build an index, in .bin format. The first...

@Aed-p Use `inverted_row_ids_map_file_path ` as example. https://github.com/myscale/MyScaleDB/blob/3bd6368796c670d56b01d83e2ba58066efb646dc/src/Storages/MergeTree/MergeTask.cpp#L397-L398 It include the full path of the disk, and `writeFile` only need to provede the file name: https://github.com/myscale/MyScaleDB/blob/3bd6368796c670d56b01d83e2ba58066efb646dc/src/Storages/MergeTree/MergeTask.cpp#L639-L640 https://github.com/myscale/MyScaleDB/blob/3bd6368796c670d56b01d83e2ba58066efb646dc/src/Storages/MergeTree/DataPartStorageOnDiskFull.cpp#L111-L121 So we should use...