Presentation is loading. Please wait.

Presentation is loading. Please wait.

Everything You Need to Know About Symphony Indexing*

Similar presentations


Presentation on theme: "Everything You Need to Know About Symphony Indexing*"— Presentation transcript:

1 Everything You Need to Know About Symphony Indexing*
*(Abridged) Joel Hahn Cooperative Computer Services

2 Overview Keyword indexes Browse indexes Analyse your indexing settings

3 Keyword indexes

4 Keyword indexing MARC Catalog records Catalog Format Policy:
Keyword Index Variations Keyword Searches Keyword Index Policies BRS Internal Synonyms BRS Internal Names (a.k.a. BRS Paragraphs) BRS Indexes

5 Keyword indexes: Variations
Where are they? Catalog Format configuration Select a format, and click Modify Go to the Entries tab Select an entry (MARC tag number), and click Modify Keyword tab

6 Keyword indexes: Variations
Editing variations Select the variation This populates the lower boxes Make any changes Click Update (or Remove) Repeat for other formats Adding variations Fill in the lower boxes Click Create

7 Keyword indexes: Variations
Parsing the display Addition/supersedes = how this variation affects the others Usually should be “In addition to others” Indicator 1 & 2 = this indexing variation will only be used if the indicators match * = any Index = which Keyword Index will be used Subfields = which subfields will be posted to the specified Index (-6 = exclude subfield 6) Leading article = how to skip A, An, The, etc. Usually should be NONE or INDICATOR2 for MARC fields

8 Keyword indexes: Variations
When do changes take effect? After the next HIR However, only for new records and records that are edited after that point The entire catalog must be reindexed for the changes to affect all existing records Does a field have to display in the OPAC in order to be indexed? No But it’s usually a good idea to include indexed fields in the FULL entries list Then searchers can discover why a record matched their search.

9 Keyword indexes: Variations
Note that the SAME, WITH, NEAR, and ADJ operators only work with data indexed according to a single keyword index variation For example: If some subfields of a field are indexed as “author”, … And other subfields are indexed as “title”, … And there isn’t any one variation that includes all of the author and title subfields, … Then a search for a word from the author portion and a word from the title portion must separate the search words with AND to get the proper results

10 Keyword indexes: Sort variations
“Sort” keyword index variations Are used to determine what data is used for sorting records by author (SAU), title (STI), or subject (SSU) in WorkFlows and e-Library If you find records that don’t sort as you’d expect, check these variations These are linked to “non-searchable” Keyword Index policies, so they aren’t truly “indexes”

11 Keyword indexes: Index policies
Parsing the display, part 1 Type Searchable = standard index You’ll mostly use this one Synonym list = group of index policies GENERAL is the only one Non searchable = can’t be searched Used only for sorting Internal synonym = link to group of BRS indexes Internal name = BRS index paragraph name where the data is indexed

12 Keyword indexes: Index policies
Parsing the display, part 2 Synonyms = list of indexes to be searched when this index is used Only used for “Synonym list” type Internal synonyms (AU, TI, etc.) are more efficient Sort index = what data is used to sort records Usually either the same as Name, or blank Use attribute = Z39.50 attribute type/number for use when people connect to your catalog via Z39.50 Can be left blank

13 Keyword indexes: Index policies
Internal name is also called “BRS paragraph” BRS is limited to 250 internal names Symphony assigns around 212 of them to keyword index policies T001-T840 are delivered for use with the indicated MARC tags AUTH, TITL, PROJ, ABS, LA, etc. are chiefly delivered for use with non-MARC records (e.g. TECHRPT, DUBLINCORE) AL01-AL36 are for textual data Numbers and text are both searchable only as text, a la 2009 {pbyr} NM01-NM06 are for numeric data Numbers are searchable as numbers, a la {pbyr} >2009 Text is still searchable as text The title-level Date Created field is automatically posted to NM01, but NM01 is not delivered assigned to an index policy; assign it to instantly get a DATECREATE index Only those not already assigned to a keyword index appear in the Internal Name list It is possible to repurpose delivered internal names, but this is best done with Client Care’s assistance

14 Keyword indexes: Index policies
The General (a.k.a. “Word or phrase”) search is a special index policy The only delivered “synonym list” type index It is automatically the sum total of all other keyword index policies, and thus all keyword index variations Except keyword indexes marked as “non-searchable” Even though it includes everything that is indexed, the SAME, WITH, NEAR, and ADJ operators still only work with data indexed according to a single keyword index variation The only way to change how a field is indexed for a General search is to change how it is indexed in all of that field’s searchable keyword index variations

15 Keyword indexes: Index policies
You can search specific keyword indexes in a General/“Word or phrase” search by enclosing the index policy name in braces terry pratchett {au} AND small gods {ti} winston churchill { } yellow submarine {505T} {pbyr} >2004<2010 Keyword indexes can also be used for creating special search forms in e-Library Such as song title, audiobook narrator, movie cast, etc.

16 Special keyword indexes: ISBN
10-digit ISBNs in the 020 $a subfield are automatically also indexed as 13-digit ISBNs If you search for a 13-digit ISBN, you will retrieve any records that contain exactly that or that only contain the 10-digit version The “extra” 13-digit ISBN does not display, it is only present in the indexes

17 Special keyword indexes: Publication year
The publication year automatically populates the PBYR Keyword Index Uses the first date found in the 008/Date1, 008/Date2, 264 field, or 260 field The same data is used for sorting by publication year Uncertain dates and ranges are indexed poorly “197u” is indexed as “197” Serials with dates are indexed (and sorted) only as “1932”

18 Special Keyword indexes: Language
Language codes in the 008 (fixed field) and 041 fields automatically populate the LANG Keyword Index If you know the 3-letter codes, you can limit searches by multiple languages at once by doing a General search for “… AND eng {lang} AND spa {lang} AND jpn {lang}” WorkFlows and e-Library use the Language policy to mask the 3-letter codes

19 Special keyword indexes: Policies
Policy fields Format, Bulletin Lists, Library, Home location, Item Type, Item Categories 1-5 Indexed in the General index as a caret, a letter, and a policy number E.g. ^F1, ^B15, ^L2, ^H436, ^T42, ^181, ^23 , ^32 , ^45 , ^51 Use the “List Policies” report to find a policy’s number Copy information is indexed in one field per unique combination of Library, Home Location, Item Type, Item Categories 1-5, and Copy shadowing Plus one field that just indexes all the owning libraries Allows more complex limits than are normally possible, by using combinations of AND, OR, NOT, SAME, etc. E.g. “Dogs AND ^B15 AND (^T42 OR ^T43) AND (^L1 NOT (^L2 or ^L3) )” Results in user-unfriendly “Your search for…” displays

20 Special keyword indexes: BRS internal synonyms
Symphony comes with six “internal synonym” lists AU, IMP, SER, SU, TI, ECOL IMP (publisher imprint) isn’t delivered searchable ECOL generally isn’t used These drive the Author, Title, Subject, and Series searches in WorkFlows and e-Library Thus, a Title search actually searches whatever fields and subfields are specified in the keyword index variations that are linked to the specified indexes Delivered settings: 130, 245, 246, 440, 730, 740, TITL, CTTL

21 Special keyword indexes: Periodical Titles
PERiodical title index Unlike Author, Title, and Subject, the Periodical/Magazine index is actually just a standard keyword index: To change which fields or subfields are indexed in PER, add or edit keyword index variations for the PER keyword index in the desired format(s)

22 Browse indexes

23 Browse Indexing MARC Catalog records Browse Index Policies
Catalog Format Policy: Browse Index Variations Browse Index Policies Browse Index Database Browse Searches MARC Authority records Authority Format Policy: Thesaurus Authority Thesaurus Policies Exact & Partial Searches Authority Searches

24 Browse indexes: Variations
Where are they? Catalog Format configuration Select a format, and click Modify Go to the Entries tab Select an entry and click Modify Browse tab

25 Browse indexes: Variations
Parsing the display Addition/supersedes = how this variation affects the others Best to use “In addition to others” Indicator 1 & 2 = this indexing variation will only be used if the indicators match * = any Index = which browse index will be used Subfields = which subfields will be posted to the specified Index (-6 = exclude subfield 6) Leading article = how to skip A, An, The Usually NONE or INDICATOR2

26 Browse indexes: Variations
Parsing the display Entry combination = used to combine authors & titles stored in separate fields into single name/title entries in the browse index For example:

27 Browse indexes: Variations
Editing variations Select the variation This populates the lower boxes Make any changes Click Update (or Remove) Repeat for other formats Adding variations Fill in the lower boxes Click Create

28 Browse indexes: Variations
When do changes or additions take effect? After the next HIR However, only for new records and records that are edited after that point The entire catalog must be reindexed for the changes to affect all existing records Does a field have to display in the OPAC in order to be indexed? Possibly not Though Display Variations are apparently used to generate the mixed-case display It’s very rare that you would want something to be browse-indexed but not display

29 Browse indexes: Exact searches
An exact search searches just the portion of the browse index generated from catalog records Does not include See and See Also references from authority records If there are any exact, character-for-character matches, those are displayed in a keyword-search-like hit list Presence or absence of subtitles can make a search fail If there are no exact matches, the system then automatically performs a standard browse search with the same search term, and those results are displayed

30 Browse indexes: Hypertext searches
If your Global Configuration > Hypertext Search Type policy is set to Partial, then when you click on a field in WorkFlows or e-Library: Any data in the subfields specified on the Hypertext tab of the Entry policy is sent as a right-truncated search of the index specified in the first Browse index variation for that Entry policy All matching headings are then joined with OR and the result is used to search the catalog

31 Browse indexes: Hypertext searches
If your Global Configuration > Hypertext Search Type policy is instead set to Exact or Keyword, then when you click on a field in WorkFlows or e-Library: Any data in the subfields specified on the Hypertext tab of the Entry policy is sent as a search of the first relevant index variation for that Entry policy Exact  Exact search in the index of the first Browse index variation Keyword  Keyword search in the index of the first Keyword Index variation

32 Analyse your indexing settings

33 Analyse your indexing settings
It is a good idea to analyze your keyword and browse indexing settings periodically Make sure everything that you want to be indexed is still indexed in the way you want it to be indexed Make sure index settings are as consistent or different between formats as they should be Make sure that any recent changes to cataloging markup and practice are reflected in the indexes

34 Analyse your indexing settings
Use the “indexparse.pl” script in the customer wiki on the Client Care web site Look for: Unicorn System Administration > Analyze Catalog Index Settings It outputs a pipe-delimited file, which can then be imported into Excel for easier analysis This data can then be summarized for librarians who want to know what fields & subfields are indexed in your catalog

35 Analyse your indexing settings
How to analyse what fields are currently included in your BRS internal synonym lists Look at the last 35 lines of Unicorn/Searchdb/KUNI/form.db on your server Or KUN0, KUN1, etc. if you have a segmented index The lists consist of BRS paragraph internal names rather than Keyword Index policy names Indexes show up in the format “T100” or “AL05” It is easy to figure out which Keyword Index policy is linked to “T100” You’ll have to look through the Keyword Index policies you’ve created to find which one is linked to “AL05”

36 Conclusion Today, we have only scratched the surface of what there is to know about Symphony indexing This should be enough to make the typical tweaks to indexing settings that librarians request Further topics include what the indexes actually look like, authority thesauri, Z39.50 searches, indexing of user & item records in BRS, and how Enterprise’s search engine works Knowing how your data is indexed is a key to becoming an expert searcher! Taking control of how your data is indexed is a key to making your catalog best serve your users!

37 Any questions?


Download ppt "Everything You Need to Know About Symphony Indexing*"

Similar presentations


Ads by Google