[Feature Request] Support Regular Expressions in File Path Matching

[yidigo]

Currently, the file query function (e.g., file()) supports basic glob patterns like * for file path matching. This is useful for selecting all files with a certain extension, but it lacks the flexibility to select files based on more complex or non-contiguous patterns.

For example, if I have a directory with daily data files named in a YYYYMMDD.parquet format, I might want to query data for specific days of the month (e.g., the 1st, 2nd, and 21st). With the current glob-only matching, my only option is to match all files and then filter them, which is inefficient.

-- This would select ALL parquet files in the directory
SELECT * FROM file('/home/data/*.parquet', Parquet);

An ideal syntax would look something like this:

-- This query should match and process only the specified files:
-- /home/data/20250901.parquet
-- /home/data/20250902.parquet
-- /home/data/20250921.parquet
SELECT * FROM file('/home/data/202509(01|02|21).parquet', Parquet);

Thank you for considering this enhancement.

[wudidapaopao]

I've checked the ClickHouse documentation, and it seems that the matching rules you mentioned are currently supported as well.
For example, use syntax similar to the following:

SELECT * FROM file('/home/data/202509{01,02,21}.parquet', Parquet);

https://clickhouse.com/docs/sql-reference/table-functions/file#globs-in-path