Write a Parquet file and verify it before returning.
Source:R/parquet_integrity.R
write_parquet_checked.RdWrapper around nanoparquet::write_parquet() that reopens the file
it just wrote and runs assert_parquet_integrity() on it. Use it
for pipeline artifacts large enough that a silent write failure
would go unnoticed; the verification is metadata-only, so it costs
milliseconds regardless of file size.
Arguments
- data
Data frame to write.
- path
Destination path.
- deep
Whether to verify by decoding every row group as well as by checking the layout. Reads the whole file back.
- ...
Passed to
nanoparquet::write_parquet().
Examples
path <- tempfile(fileext = ".parquet")
write_parquet_checked(data.frame(x = 1:10), path)
unlink(path)