I would like to come up with a standard practice to prevent any tables from having duplicates where it matters. In most cases duplicates are a combination of variables rather than one. My primary keys are just the unique ids for each field so I cannot use them. What I have been doing is querying the table first and then if the number of rows for the combination in question is 0, making the insert. However, I have read it should be possible to set up a unique key over multiple fields to enforce uniqueness. INSERT IGNORE sounds like a good possibility, however, I would need it to ignore on more than one column.
As an example, with the fields followers and followed, there can be multiple followers and followeds in a table but should only be one combination of both.
Can anyone suggest syntax first to create the unique keys over multiple fields and then to do a SQL insert query that prevents dupes? Many thanks.
To make unique index for text types you can use
(100) here after columnname tells mysql to check only first 100 bytes as unique, if you specify (4) as length then both
are same as index (both are actually
hell
in index tree) and mysql does not allow to insert them to table.See here http://dev.mysql.com/doc/refman/5.0/en/create-index.html
You can simply create a multiple-column index on these columns and enforce uniqueness: see the MySQL manual at http://dev.mysql.com/doc/refman/5.0/en/multiple-column-indexes.html.
For example, in a table with columns
id
(unique primary key),colA
andcolB
, you run:This is it: any INSERTs leading to a duplicate entry in these two columns combined will now return a MySQL error instead of going through. If you use
INSERT IGNORE
, no MySQL error will be thrown if executing it would violate this unique constraint, and yourINSERT
statement would be quietly disregarded.