Part 1:  Electronic File Transfer  (Exchange Media)

Introduction

The specifications for transferring files of MARC 21 records electronically assume the use of a standard protocol.  The File Transfer Protocol (FTP) is the most generally used.  FTP can be used in place of a physical exchange medium to move MARC 21 authority, bibliographic, classification, community information, and holdings records from a storage device in an originating system to a receiving system.  These electronic file transfer specifications were approved by the Library of Congress and the MARC 21 Advisory Committee in June 1993, and by the Canadian Committee on MARC in October 1993.

The labels used with electronic file transfer, like tape and diskette volume and file labels, communicate information from the originating system to the receiving system.  The labels identify and characterize the transaction and the information contained in the data that is being transferred electronically.  The labels will make possible an eye-readable display of the identifying information and also enable machine processing of the information.

File organization  (Electronic Transfer)

Each file of MARC 21 records should be accompanied by a companion label file containing bibliographic file and volume information.  The label and record file pair can be transferred separately or they can be compressed (archived) into a single file for transfer by the originating system.  It is then up to the receiving system to decompress (restore) the label and record file(s).  The concept of physical volumes does not apply to electronic transfers as it does to exchange using dismountable physical media, but the identification of volumes may be particularly useful with groups of files compressed for transfer into a single file.  Record blocking is not used and individual records do not span files or transfers.  Large logical files of records can be segmented into multiple physical files during separate transfers.  The following configurations may be used.

Label files  (Electronic Transfer)

A file of MARC 21 records is transferred electronically following a label file that identifies and characterizes the file of MARC 21 records, and optionally, the volume to which it belongs.  NOTE:  Volume information is not transmitted in a separate file as is done with tape and diskette.  The label file is used to transfer non-MARC information between an originating system and a receiving system.  The label file is written using only ASCII Latin characters.  A label file is required for each file of MARC 21 records transmitted.

File names  (Electronic Transfer)

These MARC 21 specifications do not mandate the length or style of file names due to differences between operating systems.  Exchange partners should agree upon file naming conventions prior to transferring MARC 21 records electronically.

Label file structure  (Electronic Transfer)

A label file consists of a sequence of fixed-length and variable fields.  The order of fields should be the same as the list below.  The length of fixed-length fields is specified in the descriptions of those fields.  Variable length fields have no maximum length.  Each field in a label file begins with a three-character tag followed by two blanks (ASCII character 20 (hex)).  The field's data content follows the two blanks.  Each field ends with a carriage return (ASCII control character 0D (hex)) or a carriage return/line feed pair (ASCII control characters 0D 0A (hex)) immediately following the data in the field.  The field terminator character (ASCII control character 1E (hex)) is no longer used.

Certain fields in the label files are specified as mandatory and must be present in each label file.  In these fields, data content or one fill character (the vertical bar | (ASCII character 7C (hex)), must be supplied.  Mandatory or optional characteristics of the fields are also noted below.  Several of the fields are repeatable.

File and volume labels - tags/fields  (Electronic Transfer)

The following order of the fields in the file label is mandatory whether or not all fields are present.

DAT (Date Compiled):  Mandatory; Fixed length; Not repeatable.  This is the date the originating system completed the compilation of the file of records.  This is not the date of the creation of the records contained in the bibliographic file.  The field is recorded according to Representation of Dates and Times (ISO 8601).  The date requires 8 numeric characters in the pattern yyyymmdd (4 for the year, 2 for the month, and 2 for the day; right justified and zero filled).  The time requires 8 numeric characters in the pattern hhmmss.f (2 for the hour, 2 for the minute, 2 for the second, and 2 for a decimal fraction of the second, including the decimal point).  The 24-hour clock (00-23) is used.

RBF (Number of Records in File):  Mandatory; Variable length; Not repeatable.  This element includes the number of logical records contained in the file of MARC 21 records.

DSN (Date Set Name):  Mandatory; Variable length; Not repeatable.  The filename of the file of MARC 21 records (which is sent separately) for which this is a file label.

ORS (Originating System ID):  Mandatory; Variable length; Not repeatable.  The name of the system that compiled the files of records.  This could be a symbol (e.g., OCLC) or text.

CID (Country Identifier):  Optional; Fixed length; Not repeatable.  The country identifier of the system that compiled the files of records.  The identifier is taken from Codes for Representation of Names of Countries (ISO 3166).

DTS (Date Sent):  Optional; Fixed length; Not repeatable.  This is the date of transmission of the file of MARC 21 records.  The field is recorded according to Representation of Dates and Times (ISO 8601).  The date requires 8 numeric characters in the pattern yyyymmdd (4 for the year, 2 for the month, and 2 for the day; right justified and zero filled).  The time requires 8 numeric characters in the pattern hhmmss.f (2 for the hour, 2 for the minute, 2 for the second, and 2 for a decimal fraction of the second, including the decimal point).  The 24-hour clock (00-23) is used.

DTR (Dates of Records):  Optional; Fixed length; Not repeatable.  This includes inclusive dates of last transaction of the records in the file, i.e. the first and last date recorded in the 005 fields of the file of records.  The field is recorded according to Representation of Dates and Times (ISO 8601).  The date requires 16 numeric characters in the pattern yyyymmddyyyymmdd (4 for the year, 2 for the month, and 2 for the day for each date; right justified and zero filled).

FOR (Format):  Mandatory; Fixed length; Not repeatable.  This element designates the format of the records, generally M for MARC.  Other codes may be defined as needed (e.g., Z for Z39.2 for copyright records).

FQF (Format Qualifier):  Optional; Variable length; Not repeatable.  This element provides additional description of the format of the record file.  For example, it may identify a particular tag set/specification for MARC records or a particular DTD for SGML records.  For MARC formats, the content of the FQF field may be text of a code from the list Z39.50 registered record syntaxes.  For DTDs, the content is the identifier in the DTD DOCTYPE element.

DES (Description of Records):  Optional; Variable length; Repeatable.  This element describes the records.  The data could be coded or describe a product name.  (For example, OCLC uses B for Bibliographic describing a data type; CDS may use a product name, such as MDS-Books All.)

CS0-n (Character Set 0-n):  Optional; Variable length; Not repeatable.  These fields specify the character sets (control and/or graphic) needed for processing the record data file.  The field content is text indicating a particular set (e.g., ISO 646-IRV, ISO Registration #37, MARC 21, or a reference to a private character set).  CS0 indicates at least the G0 set and CS1-n indicate other sets in the file.

CV0-n (Character Variation 0-n):  Optional; Variable length; Repeatable.  These fields are used in conjunction with the CS fields and contain a textual description of the variations from the set specified in the corresponding CSn field.  Variations noted in this field may be that the set:  1) was not used strictly according to the standard; 2) has options for some positions that need to be specified; or 3) has additional characters in positions that are undefined in the standard.

VOL (Volume):  Optional; Variable length; Repeatable.  This may be used if it is desirable to assign a volume number when distribution of records is by subscription.  Each file within a subscription year may be given a volume and issue number.

ISS (Issue):  Optional; Variable length; Repeatable.  This may be used if it is desirable to assign a volume and issue number when distribution of records is by subscription.  Each file within a subscription year may be given a volume and issue number.  It may be combined with Volume (e.g., V1402).

FDI (Final Destination ID):  Optional; Variable length; Not repeatable.  This field contains the name or identifier of the final destination database.

REP (Reply To):  Optional; Variable length; Repeatable.  This field contains an address given as a contact for problems/questions in transmission.  It may include an Internet or postal address.

NOT (Note):  Optional; Variable length; Repeatable.  This field contains textual information or messages about the file.

Label and record files  (Electronic Transfer of)

Each file is transferred electronically as a separate operation.  If no compression is involved, a label file should be transferred preceding the MARC 21 record file to which it relates.  If label and record files have been compressed, the transfer of all data can occur as a single operation to the receiving system.  In this case, the receiving system must process the compressed file to restore the label and record file(s) as separate files.

Example of an electronic file transfer label file:

To return, select:

Exchange Media