Skip to contents

A hash of what the table CONTAINS, independent of row order, file mtime, file count and partitioning — see the note at the top of sync.R for why byte comparison cannot do this job.

Usage

parquet_digest(path, con = NULL)

Arguments

path

a .parquet file, a directory of them, or a glob

con

optional DuckDB connection to reuse

Value

a one-row data frame: n_rows, n_groups (distinct rows), n_files, bytes, schema, schema_digest, data_digest, digest