Re: Re: How to count rows using DBI & mysql - Performance results

[Brovnik]

But I want to be able to have multiple rows with the same key/column1,
Unique keys prevent that.
--
Brovnik

[Zaxo]

hding is right. In that case, declare a multicolumn unique index. That does precisely what you are asking for. MySQL will look at the first column first, so it should be the more unique. [OB] A table needs a primary key.

create table depends
        (uses varchar(80) not null,
        usedby varchar(80) not null,
        unique deps (usedby,uses));
# naming the index is unnecessary, but may be a convenience

Btw, why not represent a leaf node by usedby=NULL? That would give a faster select for leaf/branch.

After Compline,
Zaxo

[hding]

Then why do you need to do an "INSERT ... IF NOT EXISTS"? (Note that if you're trying to enforce the uniqueness of some combination of columns, you can make the index on all of the columns; maybe this is what you mean.)

[elwarren]

In this case you should have a one to many foreign key relationship to a second table. This table would contain a unique column with all of your modules in it. Insert into this table, catch the exception if it happens.