Difference between revisions of "Dedupe 2.4"

Latest revision as of 12:20, 2 April 2013

Dedupe 2.4: Group 1 - 008 Date (Verify)

008 date

Backstage uses two date blocks from the 008 field: bytes 7-10 represent date1 and bytes 11-14 represent date2.

Verify Method

As there have been different cataloging standards through the years, sometimes it makes sense to have a more flexible method for checking dates against other records. Backstage utilizes 3 different methods (Full, Partial, Within) in an attempt to give you the kind of control you would hope to find with respect to your data.

Full

Full compares both date1 and date2 of each record; both must match:

 
 Record A:
       date1 date2
   008 1972  1973
 
 Record B:
       date1 date2
   008 1972  1973

In this example, both date1 (1972) & date2 (1973) of Record A & B match. Matching on Full returns a match for these two records.

Partial

Partial compares date1 of each record. If no match it compares date2 (as long date doesn't contain fill characters--<space> or <pipe>):

 
 Record A:
       date1 date2
   008 1971  1973
 
 Record B:
       date1 date2
   008 1972  1973

Since Record A's date1 (1971) is a mismatch against Record B's date1 (1972), both record's date2 (1973) are considered. Matching on Partial returns a match for these two records.

Within

Within compares date1 of one record to date2 of other record, and vice-versa. If no match it also does a partial verify:

 
 Record A:
       date1 date2
   008 1970  1973
 
 Record B:
       date1 date2
   008 1973  1972

In this example, date1 in Record A (1970) does not match date2 in Record B (1972). So, it checks date2 of Record A (1973) against date1 of Record B (1973). Matching on Within returns a match for these two records.

Tolerance

The date tolerance for the Verify Date parameter is the number of years the date can differ and still qualify as a matched record. This allows for more leniency with hits on the 010/020 and stricter criteria for the TITLE or other fields. The tolerance allows for dates to match within a specified number of years.

Match exactly

The year 1996 can only verify with another record that has the year 1996. It will not match either 1995, 199u, 199-, or 199 (ends with blank).

Off by 1 year

The year 1996 would match either 1995, 1996, or 1997. It would not match 1994, 1998, or years where the ending digit is missing or contains a fill character.

Off by # years

This allows the year tolerance to be increased by the Library for two-plus years. This is helpful when the desired outcome is to match the same decade of the original year in question.

• Although Backstage recommends using the 008 date blocks, if these are missing or considered unreliable, please consult with Backstage staff for using the 260 $c publication date instead.

Default

links