![]() After searching the forums, I found a post that said if your update data is exactly the same as the database data, then you get this warning ( ). I believe the UPDATE must be succeeding because there is no rollback after the statement, but there is a rollback after the INSERT.Īt a different point in my flow, I get a lot of " Warning: Output Data (22): DataWrap2ODBC: No record found" messages. I checked my MySQL logs and I see that both the UPDATE and the INSERT statement are occurring. But that is illogical, because to get a duplicate entry error, there must exist a row that was available to be updated. I continually get a duplicate entry error at this point. At one point in my flow, I am attempting to update the database using the, " Update Insert if new" output option. Check, if your DB has proper indexes and add them, if needed.I have a flow that attempts to keep a database of employee information up to date by merging data from several sources.What types of INSERT/UPDATE queries take most of server time? How can they be optimized?.Remember: select count(*) from table is much slower for innodb than for myisam.Is your application makes series of inserts to the same table in a loop? Load data infile command is much faster for series of inserts.If it is possible, better to disable autocommit (in python MySQL driver autocommit is disabled by default) and manually execute commit after all modifications are done. Is your application uses transactions? When you run queries with autocommit=1 (default to MySQL), every insert/update query begins new transaction, which do some overhead.Are any temporary tables created, especially big temporary tables?.What queries takes most of your server time?.You can catch all queries using tcpdump and maatkit Analyze your workload - catch all your queries to slowquery log and run mk-query-digest on it.key_buffer_size - very important to MyISAM, but it is used for disk temporary tables. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |