LIS 558 - MARC records

ANSI Z39.2

The various MARC (MAchine Readable Cataloging) record formats are based on the ANSI standard for bibliographic information interchange:

A record according to this standard consists of 3 parts - the leader, the directory, and the variable fields - and is terminated with ASCII code #29 (Ctrl-]).

The leader is a fixed length of 24 characters (positions 0-23), assigned as follows:
0-4 record length*
5 status
6 type of record
7 bibliographic level
8-9 [reserved]
10 indicator length*
11 identifier length* (including delimiter)
12-16 base address of data* (=leader length+directory length)
17-19 [implementation-defined]
20-23 entry map
20 length of length-of-field portion*
21 length of starting-character-position portion*
22 [length of implementation-defined portion*]
23 [reserved*] (="0")
* All numbers are expressed as base-10, using the numeric characters "0"-"9".

The directory consists of a variable number of entries and is terminated with ASCII code #30. Each entry consists of the following portions:
tag 3 characters, alphanumeric starting with "00" for a control field, with "001" reserved for a control number field and "002" for a subrecord field
length of field 0-9 characters, decimal with the length including the field terminator, and with zero indicating the maximum length with the next entry a continuation field
starting character position 0-9 characters, decimal relative to the base address of data
[implementation defined] 0-9 characters
Control field entries are ordered by tag, with numerals preceding letters; other field entries may be in any order.

In the variable field part, control fields are again ordered by tag and followed by other fields in any order. Generally, there may be more than one field with a given tag, but there is always exactly one control-number field. Each non-control ("data") field starts with a fixed-length indicator. A data element (subfield) within a non-control field may be identified by being preceded by a delimiter (ASCII code #31) plus an identifier. Control fields do not contain indicators, delimiters, or data element identifiers.

MARC records

An example of a MARC record format is the format for bibliographic data used by the Library of Congress and the National Library of Canada: There are also formats for authority data, community information, and holdings data.

Before harmonization in 1997, a slightly different bibliographic format (CanMARC) was used in Canada:

The file marcdemo.mrc contains a single sample CanMARC record. Since it contains control characters, it cannot be fully displayed directly, but, when divided into 12-character blocks, it starts with the following leader and directory:

00504nam  22
001810a 4500

The first 5 characters ("00504") in the sample record indicate, of course, that it is 504 characters long, which, in this case, is also the length of the file.

In the leader of a MARC record, the record status character is used to indicate whether the record is new or deleted; the type of record character indicates such types as printed language or printed music materials; the bibliographic level character, whether the item is a part of a monograph, a serial, or the like. Character position 8 is used for type of control (archival or not). The next 4 characters ("nam ") in the sample record can thus be interpreted as meaning "new, printed language, monograph, not archival". The MARC indicator and identifier lengths are always 2. Thus the first block of 12 characters in the sample record ends with "22".

The base address of data in the sample record can be seen to be 181 ("00181"). Character position 17 is used for encoding level (full, partial, etc.); position 18, for descriptive cataloging form (ISBD, AACR2, etc.); and position 19, for linked record requirement (yes or no). The blank at position 19, following the "0a", in the sample record indicates that no linked record is required to supply additional information about the cataloged item.

The MARC entry map is always "4500"; that is, in each entry the length-of-field portion is 4 characters, the starting-character-position portion, 5 characters; and there is no implementation-defined portion.

The directory entries can be analyzed as follows:
Tag Length of field Starting character position
001 0013 00000
008 0042 00013
016 0017 00055
020 0010 00072
040 0024 00082
055 0011 00106
082 0012 00117
100 0030 00129
245 0053 00159
260 0042 00212
300 0033 00254
650 0016 00287
650 0019 00303
The 1999 MARC bibliographic format has control fields 005 for date and time of latest transaction and 010 for Library of Congress control number, but these are not present in the sample (unharmonized) CanMARC record.

If we take the 30 characters starting at position 181+129=310, we shall have the contents of the 100 field, the main entry. Representing ASCII codes #30 and #31 with and § respectively, this appears as

1 §aPilcher, F. E. V.,§d1912-¶

If linked records are required for a MARC record, they can be identified in a subfield of the appropriate linking entry field. The format commonly used is the control number of the linked item preceded, in parentheses, by the MARC code for the agency to which the control number applies (e.g, "(CaOONL)").


Last updated July 5, 2001.
This page maintained by Prof. Tim Craven
E-mail (text/plain only):
Faculty of Information and Media Studies
University of Western Ontario,
London, Ontario
Canada, N6A 5B7