What this MARC 21 inspector validates
The parser treats each record as bytes, not lines. It checks the 24-byte Leader, declared logical record length, ASCII 1D record terminator, base address, the MARC 21 4500 entry map, 12-byte Directory entries, numeric field lengths and starting positions, field boundaries, overlaps, gaps, and ASCII 1E field terminators. It then separates control and data fields, reads the two indicators and ASCII 1F subfield delimiters, and preserves exact byte offsets in the CSV audit.
Batch-level checks
The record directory exposes Leader/06 type, Leader/07 bibliographic level, Leader/09 encoding, field counts, titles, primary names, and control numbers. Duplicate field 001 values are flagged across the batch. Likely bibliographic records without field 245, duplicate or missing 001 fields, and non-40-byte field 008 values remain visible. These are structural and high-level completeness checks, not a full validation profile.
UTF-8 and MARC-8 handling
Leader/09 code a is decoded as strict UTF-8 and invalid byte sequences are errors. A blank Leader/09 declares MARC-8. Because MARC-8 uses combining characters, escape sequences, and multiple coded character sets, this page preserves non-ASCII MARC-8 bytes as visible \xNN escapes instead of mislabeling them as another encoding. Use a qualified MARC-8 converter when readable character conversion is required.
Cataloging and import boundary
A structurally clean file is not necessarily a valid, complete, or import-safe catalog record. This tool does not implement every MARC 21 format, field-specific indicator/code list, RDA/AACR2 rule, local profile, authority control, linked-data requirement, ILS item/holdings mapping, normalization, or MARC-8 conversion. Validate against the exact record format and receiving system, inspect a representative sample, retain the source batch, and use a mature cataloging tool such as MarcEdit for controlled correction.
Frequently asked questions
Can I open a .mrk text file?
No. This page expects ISO 2709 binary records with Leaders, Directories, field terminators, and record terminators. Mnemonic text, MARCXML, and MARC-in-JSON need different parsers.
Why are MARC-8 accents shown as byte escapes?
Showing raw bytes is safer than pretending they are Windows-1252 or Latin-1. Correct MARC-8 display requires its character-set tables, escape sequences, and combining-character rules.
Does a missing 245 always mean the record is invalid?
No. The warning is limited to records whose Leader/06 resembles a bibliographic format, and local or specialized workflows can differ. Apply the correct MARC format and profile before deciding.
Can the inspector repair a damaged batch?
No. It can recover some record boundaries for diagnosis, but it never writes a repaired MARC file. Automatic repair could change field offsets, encoding, or catalog data without enough context.
Authoritative format references
The byte model follows the Library of Congress MARC 21 record-structure specification, which defines the Leader, 12-byte Directory entries, field terminator, and record terminator. Leader positions and bibliographic codes follow the MARC 21 bibliographic Leader documentation; character-handling boundaries follow the MARC 21 character-set specification.
Inspect a COBOL copybook · Inspect EBCDIC bytes · Browse legacy-data tools