Re: distinct clause issue for 600000 records

[Øystein Grøvlen]

Hi,

Here are several things note:

1. LEFT JOIN: Don't use it if you do not mean it.

LEFT JOIN implies "optional". That is, for each row in the left hand table there might or might not be information in the right hand table. LEFT JOIN is used to make sure rows from the left hand table is included in the result even when there is no match in the right hand table. Example: We want to list persons in person table even if there is no registered address in address table.

Unnecessary use of LEFT JOIN restricts optimizer in what query plan it may use since left hand tables needs to be processed before right hand tables. In other words, your query only allows for one join order.


2. ON clause is for join conditions

Normally single table conditions in ON clause does not make much sense. Take for example your condition "ea2.address_type='Bil'. If using LEFT JOIN this will not filter out rows with other address_types. Those rows will still be included, but ea2 columns will be NULL instead of actual data. Is that what you want? If not, you should put this criteria in the WHERE clause instead.


3. Do not include tables that are not used in query.

In your case, the only effect of tables beship, beinv, and reginv is that they may generate duplicate rows in your result set.


Hope this helps,

Øystein Grøvlen,
Senior Principal Software Engineer,
MySQL Group, Oracle,
Trondheim, Norway