bibxml2

A simple converter of (possibly compressed) MARCXML/PICAXML to (possibly compressed) CSV/TSV/parquet.

The resulting CSV/TSV/parquet has been designed to be easy to use as a data table, but also to retain all ordering informaation in the original when such is needed. The format is as follows: record_number,field_number,subfield_number,field_code,subfield_code,value

Here, record_number identifies the MARC/PICA+ record, while field_number and subfield_number can be used for more exact filtering / reconstructing the original field structure/order if needed.

For MARC data fields, ind1 and ind2 values are reported as separate rows with the subfield_code being Y or Z, but only when non-empty (MARC requires subfield codes to be lowercase, so this should be relatively safe). The MARC leader is output with field code LDR.

Installation

Install from pypi with e.g. pipx install bibxml2.

Usage

Usage: marcxml2 [OPTIONS] [INPUT]...

  Convert from MARCXML (compressed) input files into (compressed) CSV/TSV/parquet

Options:
  -o, --output TEXT  Output CSV/TSV (compressed) / parquet file  [required]
  --help             Show this message and exit.

Usage: picaxml2csv [OPTIONS] [INPUT]...

  Convert from PICAXML (compressed) input files into (compressed) CSV/TSV/parquet

Options:
  -o, --output TEXT  Output CSV/TSV (compressed) / parquet file  [required]
  --help             Show this message and exit.

If the output file extension is .parquet, the output will be in parquet format, compressed with zstd, and with field typings maximally compatible with common R and Python ecosystems. Otherwise, compressed files will be read/written if the filename ends with an identifier recognised by fsspec. TSV format will be used if the output filename contains .tsv, otherwise CSV will be used.