MARC data structure

This note discusses the MARC record of a work published in subscription form in 1837 - 1839 called "American Scenery." I made a website about this work for Information Organization, LIS-5703. For a discussion of "American Scenery" in terms of LCC, DDC, LCSH, AAT, ADL, AACR2, DC, and EAD, I refer you to that earlier work.

Three different views of the same MARC record are presented, courtesy of Orbis, the Yale University Library catalog. The raw view is completely unformatted. It is not exactly byte for byte correct (some control characters got mangled in the various cut and paste operations), but it is very close to what a computer would see processing a MARC record. The staff view is a traditional librarian or cataloguer view of a MARC record. The patron view shows how MARC data can be organized for presentation to information seekers.

After showing the three views, I discuss the data structure of the MARC record in some detail.

Raw View

In "raw" form (again, slightly altered by copying into a text editor):

02617cam 22004931a 450000100080000000500170000800800410002503500230006603500140008904000170010304300120012009000210013210000420015324503360019 524600210053124600630055226000350061530000440065059001610069459001990085550000190105450001420107350000510121550000390126650 001350130550001460144065100430158665100360162965500470166565500630171269200280177569200250180369200290182870000500185770000 320190770000300193970000340196970000340200370000190203770000220205692700090207892800130208794800230210011974352003022713 0037.0940923s1840 enkabcf 00 0 eng u a(UPRA)CTYXRL7078-B 9CAF1680YL cUPRAdCtY-BR an-us--- a\Za W679\b+840s1 aWillis, Nathaniel Parker,d1806-1867.10aAmerican scenery, or, Land, lake, and river illustrations of transatlantic nature :buniform with Dr. Beattie's Switzerland, Scotland, & Waldenses /cfrom drawings by W.H. Bartlett, engraved in the first style of the art, by R. Wallis, J. Cousen, Willmore, Brandard, Adlard, Richardson, &c ; the literary department by N.P. Willis.30aAmerican scenery30aLand, lake and river illustrations of transatlantic nature aLondon :bGeorge Virtue,c1840 a30 parts :bills., map, port. ;c29 cm. aBEIN Za W679 +840S: Original wrappers. On parts 1-15, price ([50] cents) has been stamped on. Without subscriber list in part 1. Bookplate of Owen F. Aldis. aBEIN 2003 +56: Original wrappers. Variant wrapper for part 1: advertising matter on p. [2-3]. Plates are hand-colored, probably in the 20th century. Stamp: Carl Gustaf. Bookplate of Paul Mellon. aWrapper title. a"The first edition of this work is printed expressly for Martin and Co., 42, Whitehall Street, New York, and at Philadelphia and Boston." a"Subscriber's copy. Sold to subscribers only." aPrinted subscriber list in part 1. aAdded engraved title pages in parts 1 and 30: American scenery by N.P. Willis ; illustrated in a series of views by W.H. Bartlett. aFor possible binding into two volumes, title pages (both dated 1840) and lists of engravings for vol. 1 and vol. 2 included at end of pt. 30. 0aUnited StatesxDescription and travel. 0aUnited StatesvPictorial works. 7aHand-colored illustrations.2local5CtY-BR 7aPublishers' advertisementszEnglandzLondony1840.2rbgenr14aAldis, OwenxBookplate.14aGustaf, CarlxStamp.14aMellon, PaulxBookplate.1 aBartlett, W. H.q(William Henry),d1809-1854.1 aWallis, Robert,d1794-1878.1 aCousen, John,d1804-1880.1 aWillmore, Arthur,d1814-1888.1 aBrandard, Robert,d1805-1862.1 aAdlard, Henry.1 aRichardson, G. K. arlin aAC031397 aRLA Beinecke Recon

Staff (MARC) View

Orbis presents a staff view that is much more legible than the raw dump:

000 02617cam 22004931a 450
001 1197435
005 20030227130037.0
008 940923s1840 enkabcf 00 0 eng u
035 __ |a (UPRA)CTYXRL7078-B
035 __ |9 CAF1680YL
040 __ |c UPRA |d CtY-BR
043 __ |a n-us---
090 __ |a \Za W679\ |b +840s
100 1_ |a Willis, Nathaniel Parker, |d 1806-1867.
245 10 |a American scenery, or, Land, lake, and river illustrations of transatlantic nature : |b uniform with Dr. Beattie's Switzerland, Scotland, & Waldenses / |c from drawings by W.H. Bartlett, engraved in the first style of the art, by R. Wallis, J. Cousen, Willmore, Brandard, Adlard, Richardson, &c ; the literary department by N.P. Willis.
246 30 |a American scenery
246 30 |a Land, lake and river illustrations of transatlantic nature
260 __ |a London : |b George Virtue, |c 1840
300 __ |a 30 parts : |b ills., map, port. ; |c 29 cm.
590 __ |a BEIN Za W679 +840S: Original wrappers. On parts 1-15, price ([50] cents) has been stamped on. Without subscriber list in part 1. Bookplate of Owen F. Aldis.
590 __ |a BEIN 2003 +56: Original wrappers. Variant wrapper for part 1: advertising matter on p. [2-3]. Plates are hand-colored, probably in the 20th century. Stamp: Carl Gustaf. Bookplate of Paul Mellon.
500 __ |a Wrapper title.
500 __ |a "The first edition of this work is printed expressly for Martin and Co., 42, Whitehall Street, New York, and at Philadelphia and Boston."
500 __ |a "Subscriber's copy. Sold to subscribers only."
500 __ |a Printed subscriber list in part 1.
500 __ |a Added engraved title pages in parts 1 and 30: American scenery by N.P. Willis ; illustrated in a series of views by W.H. Bartlett.
500 __ |a For possible binding into two volumes, title pages (both dated 1840) and lists of engravings for vol. 1 and vol. 2 included at end of pt. 30.
651 _0 |a United States |x Description and travel.
651 _0 |a United States |v Pictorial works.
655 _7 |a Hand-colored illustrations. |2 local |5 CtY-BR
655 _7 |a Publishers' advertisements |z England |z London |y 1840. |2 rbgenr
692 14 |a Aldis, Owen |x Bookplate.
692 14 |a Gustaf, Carl |x Stamp.
692 14 |a Mellon, Paul |x Bookplate.
700 1_ |a Bartlett, W. H. |q (William Henry), |d 1809-1854.
700 1_ |a Wallis, Robert, |d 1794-1878.
700 1_ |a Cousen, John, |d 1804-1880.
700 1_ |a Willmore, Arthur, |d 1814-1888.
700 1_ |a Brandard, Robert, |d 1805-1862.
700 1_ |a Adlard, Henry.
700 1_ |a Richardson, G. K.
927 __ |a rlin
928 __ |a AC031397
948 __ |a RLA Beinecke Recon

Patron View

Finally, in patron-friendly form it is presented as:

Author: Willis, Nathaniel Parker, 1806-1867.

Title: American scenery, or, Land, lake, and river illustrations
of transatlantic nature : uniform with Dr. Beattie's
Switzerland, Scotland, & Waldenses / from drawings by
W.H. Bartlett, engraved in the first style of the art,
by R. Wallis, J. Cousen, Willmore, Brandard, Adlard,
Richardson, &c ; the literary department by N.P. Willis.

American scenery
Land, lake and river illustrations of transatlantic nature

Published: London : George Virtue, 1840

Description: 30 parts : ills., map, port. ; 29 cm.

Location: BEINECKE (Non-Circulating)

Call Number: 2003 +56

Library has: pt.1-pt.30
______________________________

Location: BEINECKE (Non-Circulating)

Call Number: Za W679 +840s

Library has: pt.1-pt.30
______________________________

Location: MUDD LIBRARY, Zeta Collection (Non-Circulating)

Call Number: E165 W55 (LC)+ Oversize

Status: Not Checked Out

Variant and related titles:
American scenery
Land, lake and river illustrations of transatlantic nature

Local notes: BEIN Za W679 +840S: Original wrappers. On parts 1-15, price
([50] cents) has been stamped on. Without subscriber
list in part 1. Bookplate of Owen F. Aldis.
BEIN 2003 +56: Original wrappers. Variant wrapper for part
1: advertising matter on p. [2-3]. Plates are
hand-colored, probably in the 20th century. Stamp: Carl
Gustaf. Bookplate of Paul Mellon.

Notes: Wrapper title.
"The first edition of this work is printed expressly for
Martin and Co., 42, Whitehall Street, New York, and at
Philadelphia and Boston."
"Subscriber's copy. Sold to subscribers only."
Printed subscriber list in part 1.
Added engraved title pages in parts 1 and 30: American
scenery by N.P. Willis ; illustrated in a series of
views by W.H. Bartlett.
For possible binding into two volumes, title pages (both
dated 1840) and lists of engravings for vol. 1 and vol.
2 included at end of pt. 30.

Subjects (Library of Congress):
United States Description and travel.
United States Pictorial works.

Subjects (Local Yale):
Aldis, Owen
Gustaf, Carl
Mellon, Paul

Type of Material: Hand-colored illustrations.
Publishers' advertisements England London 1840.

Also listed under:
Bartlett, W. H. (William Henry), 1809-1854.
Wallis, Robert, 1794-1878.
Cousen, John, 1804-1880.
Willmore, Arthur, 1814-1888.
Brandard, Robert, 1805-1862.
Adlard, Henry.
Richardson, G. K.

Data Structure

The data structure of a MARC record is described in various publications of the Library of Congress. This description is based on the 2002 Concise Edition of MARC 21 Concise Format for Bibliographic Data at http://www.loc.gov/marc/bibliographic/ecbdhome.html. There seem to be dozens of specifications, all closely related, and I'm not an expert in the differences. As described in this source, every MARC record consists of three parts.

Leader

The first part of a MARC record is the "Leader", a 24 byte fixed length field containing processing data about the entire record. For instance, bytes 00-04 (five bytes) specify the length of the entire record. Bytes 12-16 are the base address of the actual data. Other bytes in the leader specify various status conditions, types, modes, character sets, etc, each in their appointed slot.

Looking at the "raw" record for "American Scenery," the first twenty four bytes ("02617cam 22004931a 450") are the leader. 02617 is the length, etc. If you look at bytes 12-16 you see 00493 which is the byte count where the first variable record starts. In the staff view , this part of the record is dentified with a tag of "000". In the patron representation of a MARC record this information is missing, as it is not important to a patron.

Directory

The second part of the MARC data structure is the "Directory". The directory starts in character position 24 and contains a series of twelve byte index entries where each index entry consists of a three byte type field, a four byte length field, and a five byte starting position field.

Again, the "raw" view of the "American Scenery" record shows us the data:

000100080000000500170000800800410002503500230006603500140008904000170010304300120012009000210013210000420015324503360019524 600210053124600630055226000350061530000440065059001610069459001990085550000190105450001420107350000510121550000390126650001 350130550001460144065100430158665100360162965500470166565500630171269200280177569200250180369200290182870000500185770000320 1907700003001939700003401969700003402003700001902037700002202056927000902078928001302087948002302100

A little rearrangement (is it information now?) shows that this is actually 39 twelve byte fields:

001000800000
005001700008
008004100025
035002300066
035001400089
040001700103
043001200120
090002100132
100004200153
245033600195
246002100531
246006300552
260003500615
300004400650
590016100694
590019900855
500001901054
500014201073
500005101215
500003901266
500013501305
500014601440
651004301586
651003601629
655004701665
655006301712
692002801775
692002501803
692002901828
700005001857
700003201907
700003001939
700003401969
700003402003
700001902037
700002202056
927000902078
928001302087
948002302100

As stated before, each of these twelve byte fields has three parts:

So, to pick one at random, "651003601629" is MARC type 651, is 36 bytes long, and starts at byte 1629. The data it points to is "|a United States |v Pictorial works." which is indeed 36 bytes long.

If you look at the first three bytes of each index entry, you'll see the MARC tags: the two 035's, etc. These are also reflected in the staff view.

Finally, one other observation. If you count the number of records records under the "Staff" view, there are forty. However, the MARC record only has 39. The staff view of the leader in a "000" record is a fabrication!

Data fields

The third part of the MARC data structure consists of the data fields. These are variable length records indexed by Directory entries. They are essentially pure data; structure is provided in the Directory.

The "American Scenery" MARC data in raw form:

s1840 enkabcf 00 0 eng u a(UPRA)CTYXRL7078-B 9CAF1680YL cUPRAdCtY-BR an-us--- a\Za W679\b+840s1 aWillis, Nathaniel Parker,d1806-1867.10aAmerican scenery, or, Land, lake, and river illustrations of transatlantic nature :buniform with Dr. Beattie's Switzerland, Scotland, & Waldenses /cfrom drawings by W.H. Bartlett, engraved in the first style of the art, by R. Wallis, J. Cousen, Willmore, Brandard, Adlard, Richardson, &c ; the literary department by N.P. Willis.30aAmerican scenery30aLand, lake and river illustrations of transatlantic nature aLondon :bGeorge Virtue,c1840 a30 parts :bills., map, port. ;c29 cm. aBEIN Za W679 +840S: Original wrappers. On parts 1-15, price ([50] cents) has been stamped on. Without subscriber list in part 1. Bookplate of Owen F. Aldis. aBEIN 2003 +56: Original wrappers. Variant wrapper for part 1: advertising matter on p. [2-3]. Plates are hand-colored, probably in the 20th century. Stamp: Carl Gustaf. Bookplate of Paul Mellon. aWrapper title. a"The first edition of this work is printed expressly for Martin and Co., 42, Whitehall Street, New York, and at Philadelphia and Boston." a"Subscriber's copy. Sold to subscribers only." aPrinted subscriber list in part 1. aAdded engraved title pages in parts 1 and 30: American scenery by N.P. Willis ; illustrated in a series of views by W.H. Bartlett. aFor possible binding into two volumes, title pages (both dated 1840) and lists of engravings for vol. 1 and vol. 2 included at end of pt. 30. 0aUnited StatesxDescription and travel. 0aUnited StatesvPictorial works. 7aHand-colored illustrations.2local5CtY-BR 7aPublishers' advertisementszEnglandzLondony1840.2rbgenr14aAldis, OwenxBookplate.14aGustaf, CarlxStamp.14aMellon, PaulxBookplate.1 aBartlett, W. H.q(William Henry),d1809-1854.1 aWallis, Robert,d1794-1878.1 aCousen, John,d1804-1880.1 aWillmore, Arthur,d1814-1888.1 aBrandard, Robert,d1805-1862.1 aAdlard, Henry.1 aRichardson, G. K. arlin aAC031397 aRLA Beinecke Recon

From a data structure point of view, the data area is very simple. It is simply variable length strings concatenated into a block.

Summary

This is probably more than you ever wanted to know about the data structure of a MARC record, but it is pretty interesting the way the pieces all fit together.

To summarize, a MARC data structure is a variable length structure consisting of three parts: a fixed length header, a variable number of fixed length index records, and a corresponding number of variable length data fields.