How to configure diacritics / special characters for index and search in Primo?
- Article Type: Q&A
- Product: Primo
- Product Version: 4
Question
How to configure that a search for regular Latin letter will also search special character (e.g. Umlaut), and the opposite?
Answer
There are 3 levels of normalization for special characters in Primo:
1. The lowest are legacy mappings that are hardcoded in Java.
2. The second are system-defined OTB files. They are located in /exlibris/primo/p4_1/ng/jaguar/home/profile/analysis/specialCharacters. (for example non_cjk_unicode_normalization.txt and all other txt files in that directory besides user_defined_unicode_normalization.txt).
3. The third and the highest is user_defined_unicode_normalization.txt file. Each upper level supersedes the lower ones.
To configure search an index for special characters please do the following:
1. On all FEs and SEs server, edit user_defined_unicode_normalization.txt the following way:
se_profile
cd analysis/specialCharacters
vi user_defined_unicode_normalization.txt
Add codes in the following pattern:
<Unicode> TAB <Unicode> (meaning Unicode=Unicode).
e.g. if you want ? to be searchable for s, you’ll have to insert 0161 TAB 0073 -> ?=s
Please note that you have to use TAB and not spaces.
2. Deploy all
3. Backup current indexes
4. Run clean_indexes.sh from be_bin
5. Run indexing_and_didumean_and_hotswapping process
Additional Information
The above instructions do not affect the sort order in Primo.
Category: BO Configuration
Subject: Code & Mapping Tables
- Article last edited: 11/21/2013