Dataframes powered by a multithreaded, vectorized query engine, written in Rust
29.49k
stars
1.87k
forks
source link
`read_ndjson()` and `read_parquet()` behave differently when the input is a list of files with different schemas #18306
Open
etiennebacher opened 1 month ago
Checks
Reproducible example
Log output
Issue description
read_ndjson()
andread_parquet()
behave differently when the input is a list of files with different schemas:read_ndjson()
only keeps the schema of the first file and adds empty rowsread_parquet()
ignores (probably the best behavior here)Expected behavior
Both functions should have the same behavior.
Installed versions