# SOT SQL question **URL:** \ **Category:** Rails\ **Created:** [August 16, 2008, 9:53pm UTC](https://www.ruby-forum.com/t/sot-sql-question/146055 "2008-08-16T21:53:29Z")\ **Posts on this page:** 5\ **Page:** 1
**Author:** ![Jeffrey\_LSTaylor](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/j/ee7513/32.png) [@Jeffrey\_LSTaylor](https://www.ruby-forum.com/u/Jeffrey_LSTaylor)\ **Post date:** [August 16, 2008, 9:53pm UTC](https://www.ruby-forum.com/t/sot-sql-question/146055/1 "2008-08-16T21:53:29Z")
I have an items tables, a tokens table, and a items\_tokens join table. There can be multiple joins for an items, tokens pair, i.e. a token appears multiple times in an item. I would like to increment a column in the tokens for EACH time it appears in in the join table. It seems that “UPDATE tokens JOIN items\_tokens ON items\_tokens.token\_id = tokens.id set occurrences = occurrences + 1 WHERE items\_tokens.item\_id = 339;” (for item 339) would do the trick but it only increments each occurrences at most once, i.e. duplicates are removed before incrementing occurrences. Is there a way to do this in ONE SQL statement? TIA, Jeffrey ---
**Author:** ![Jeffrey\_LSTaylor](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/j/ee7513/32.png) [@Jeffrey\_LSTaylor](https://www.ruby-forum.com/u/Jeffrey_LSTaylor)\ **Post date:** [August 17, 2008, 7:15pm UTC](https://www.ruby-forum.com/t/sot-sql-question/146055/2 "2008-08-17T19:15:27Z")
Why don’t you just use a counter\_cache on Token? ---
**Author:** ![Jeffrey\_LSTaylor](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/j/ee7513/32.png) [@Jeffrey\_LSTaylor](https://www.ruby-forum.com/u/Jeffrey_LSTaylor)\ **Post date:** [August 18, 2008, 5:24am UTC](https://www.ruby-forum.com/t/sot-sql-question/146055/3 "2008-08-18T05:24:54Z")
ActiveRecord was great for prototyping the proof of concept. But it’s too slow for production. I need to keep it in all SQL. Thanks, Jeffrey Quoting Rein H. [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid): ---
**Author:** ![Jeffrey\_LSTaylor](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/j/ee7513/32.png) [@Jeffrey\_LSTaylor](https://www.ruby-forum.com/u/Jeffrey_LSTaylor)\ **Post date:** [August 18, 2008, 9:56pm UTC](https://www.ruby-forum.com/t/sot-sql-question/146055/4 "2008-08-18T21:56:02Z")
Sorry, I wasn’t specific enough. It is too slow for one of my applications for just one of the models w/ a join table. When a cronjob that is invoked every five minutes occasionally runs for over 20 minutes with actual data, an application is too slow and it is time to start profiling and benchmarking. The results are interesting, to me at least. Pure Active Record: 7.2sec SQL w/ join table: 0.77sec SQL w/out join table: 0.20sec The application is an adaptive RSS reader. Every article has many words w/ statistics for each word (occurrences, click-thrus, up-votes, down-votes, etc.). It is faster to regenerate the word list on the fly each time and select on the word’s string field than use a join table with select with joins on the IDs. As always, YMMV. So measure, then cut (code). Jeffrey Quoting Rein H. [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid): ---
**Author:** ![Jeffrey\_LSTaylor](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/j/ee7513/32.png) [@Jeffrey\_LSTaylor](https://www.ruby-forum.com/u/Jeffrey_LSTaylor)\ **Post date:** [August 18, 2008, 6:24am UTC](https://www.ruby-forum.com/t/sot-sql-question/146055/5 "2008-08-18T06:24:31Z")
LMAO. Good luck with that. And don’t tell all the people running Rails with AR in production. I’m sure they’ll be devastated to hear that their sites are too slow. On Aug 17, 11:23 pm, “Jeffrey L. Taylor” [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid)