Subject: | Re: Not quie sure how to do this
| Date: | Mon, 20 Jun 2016 12:28:56 +1000
| From: | Tom Krieg <"tom krieg AT outlook.com">
| Newsgroups: | pnews.paradox-client_server
|
In fact, with the inner join and all the outer joins required to get the
data I need, and an 'order by' on 3 fields the query hasn't taken more
than 860ms. This is great because by the time the user's finished typing
the query has completed. It's a "type-ahead" query on address.
On 20/06/2016 11:54 AM, Tom Krieg wrote:
> I joined a couple of fields into a single index and tried your
> suggestion below, and got the query down to 860ms which is more than
> acceptable. Thanks Liz.
>
> On 19/06/2016 11:21 PM, Liz McGuire wrote:
>>
>> SELECT d.*
>> FROM adb_street s
>> INNER JOIN adb_address_detail d
>> ON d.adb_address_detail.street_uid = s.adb_street.street_uid
>> WHERE s.adb_street.street_name like 'THEOD%'
>>
>> ...and that video may have other ideas for you.
>>
>> You could try an index with street_uid first and street_name second, to
>> see if that will speed matters - I'm just not sure how MS SQL Server
>> uses indexes, so I don't know if they key and a separate index on
>> street_name will work, or if both need to be in one index, or what. This
>> is why I keep telling the bosses we need a DBA.
>>
>> Liz
>>
--
To send me an email, remove the spaces.
|