Details
Description
The following snippet of code:
<unique name="NAME_IDX">
<unique-column name="NAME"/>
</unique>
Schema:
<!ELEMENT index (option*,index-column+)>
<!ATTLIST index
name CDATA #IMPLIED
generates the following mysql code (the name is totally ignored), but the schema allows it.
Create ...
...
UNIQUE (NAME)
);
The unique index name (NAME_IDX) is missing, it should correctly be:
Create ...
UNIQUE NAME_IDX (NAME);
);
changing the unique.vm in sql/base/mysql/unique.vm from
#foreach ($unique in $table.Unices)
UNIQUE($unique.ColumnList),
#end
to
#foreach ($unique in $table.Unices)
UNIQUE $unique.Name ($unique.ColumnList),
#end
I think we have to change unique.vm to do the following semantics:
if ("unique name" is specified) {
unique $uniq.name ('column')
}
else {
unique ('column')
}
We could also set the name whenever creating an sql unique column For example, if you have to maintain two different databases (i.e. oracle and mysql), both generated from the same xml file, you have probably written a tool which compares both databases. Oracle always uses an unique index name (afaik), mysql can do without, but comparisson from your tool failes.