Question: What does 'scan indexes' on the advanced search page mean?
Answer: When you choose an index, enter a term, click 'scan indexes' and do the search, Koha displays the searched term and the following terms found in this index with the number of corresponding records That is search is not made directly in the catalog, but first in the indexes It works only for one index at once, and only with no limit in Location (All libraries needed)
Question:What is the difference between a keyword search using the '*' (asterisk) versus a keyword search using the '%' (percent)? Both work in the catalog, but return different sets. Why?
Answer: A wildcard is a character (*,?,%,.) that can be used to represent one or more characters in a word. Two of the wildcard characters that can be used in Koha searches are the asterisk ('*') and the percent sign ('%'). However, these two characters act differently when used in searching.
The '*' is going to force a more exact search of the first few characters you enter prior to the '*'. The asterisk will allow for an infinite number of characters in the search as long as the first few characters designated by your search remain the same. For example, searching for authors using the term, Smi*, will return a list that may include Smith, Smithers, Smithfield, Smiley, etc depending on the authors in your database.
The '%' will treat the words you enter in the terms of "is like". So a search of Smi% will search for words like Smi. This results in a much more varied results list. For example, a search on Smi% will return a list containing Smothers, Smith, Smelley, Smithfield and many others depending on what is your database.
The bottom line in searching with wildcards: '*' is more exact while '%' searches for like terms.
Question: Why does my Zebra title search for 'Help' not turn up 'The help' in the first pages of results?
Answer: When doing a title search, you actually want to search for the title (i.e., 'the help' rather than just 'help'), and it will bubble right up to the top. If you're just searching for 'help' then the relevance ranking is going to affect the results you see.
When it comes to relevance in Zebra, here's what's happening. First, the search is done. If you search for the title "help", then any title that has "help" in it comes back. Then from those records, separately, it does relevance on the *whole* record. The more your word appears, the more relevant, and some MARC tags are worth more points than others. So a self-help book with 505 notes where "help" appears a *lot* will be at the top, regardless of keyword or title.
But when you add a *second word*, that helps it figure things out, as it's weighing the relevance of both words and the phrase. Because of the way relevance works, if you search "the help", then "the help" or "the help I need" are more relevant than "the way to help", because they appear together in order. Likewise, "help the girl" would be lower relevance, because it's out of order, and "help for the homeless" would be lower still, as they're out of order, and apart.
The moral of the story is that single word searches, particularly on common words, will always struggle a bit; it can't evaluate relevance well, because you've not given it enough input.