yuriyandropov
Active member
I will try to gather in this thread some thoughts on how to improve the indexxx. Everybody: feel free to join the debate.
1. SQL Search. This is arguably what differentiates indexx from other sites, like ebi or thenude.eu.
I can search both set names and model names from the same form. From what I can surmise, the query uses SQL WHEN ... LIKE clause including the the _ and % wildcards. This is very good first step. For example I can type "062_j%" and get back DDF sets "062bj, 062fjh, 062bjfjh", but not, for example "062hot", as it does not contain in the 5th position in the string. So far so good. The problem is when I get a lot of entries back I can not sort them by criteria of my choosing.
I would also like gradual implementation of search using simplified regular expressions. Most SQL DB engines support XQuery by now, for example MySQL support it as of ver. 5.0
2. I am not sure how your "Similar names found" algorithm is implemented. As it is now, it gives back a lot of useless data. For example, using the same search string "062_j%" return a lot of model names beginning with "J". This is way off target. At the same time SOUNDEX ( "phonetic match") is probably not implemented. When I type in "Olexandra" ( Ukrainian spelling of Alexandra"), the engine does not return any "Alexandra" from the index, as it should.
1. SQL Search. This is arguably what differentiates indexx from other sites, like ebi or thenude.eu.
I can search both set names and model names from the same form. From what I can surmise, the query uses SQL WHEN ... LIKE clause including the the _ and % wildcards. This is very good first step. For example I can type "062_j%" and get back DDF sets "062bj, 062fjh, 062bjfjh", but not, for example "062hot", as it does not contain in the 5th position in the string. So far so good. The problem is when I get a lot of entries back I can not sort them by criteria of my choosing.
I would also like gradual implementation of search using simplified regular expressions. Most SQL DB engines support XQuery by now, for example MySQL support it as of ver. 5.0
2. I am not sure how your "Similar names found" algorithm is implemented. As it is now, it gives back a lot of useless data. For example, using the same search string "062_j%" return a lot of model names beginning with "J". This is way off target. At the same time SOUNDEX ( "phonetic match") is probably not implemented. When I type in "Olexandra" ( Ukrainian spelling of Alexandra"), the engine does not return any "Alexandra" from the index, as it should.