On 24 Aug 2009, at 5:05am, Tito Ciuro wrote:
> What I meant is that if a table contains several columns, some or all
> of which need to be searched using 'ends with', then replicating the
> columns (by reversing the string) and keeping extra indices could
> potentially affect performance
Hi Simon,
On Aug 23, 2009, at 8:52 PM, Simon Slavin wrote:
>
> On 24 Aug 2009, at 4:22am, Tito Ciuro wrote:
>
>> But that would introduce the overhead of doubling the space required
>> for every string + an additional column index.
>
> One of the options I mentioned was to store the reversed
On 24 Aug 2009, at 4:22am, Tito Ciuro wrote:
> But that would introduce the overhead of doubling the space required
> for every string + an additional column index.
One of the options I mentioned was to store the reversed string and
not the original. If you need to reconstruct the original
Hi Simon,
But that would introduce the overhead of doubling the space required
for every string + an additional column index. If the schema contains
more columns where this type of query needs to take place, it seems to
me that this would not be a good solution.
Thanks for your help,
--
On 24 Aug 2009, at 3:44am, Tito Ciuro wrote:
> Is there a way to optimize this type of queries? (column Value is
> indexed):
>
> SELECT Value FROM MyValues WHERE Value LIKE '%crashed.'
>
> I've seen the document where 'begins with' queries can be optimized
> using >= and < (end of the '4.0 The
Hello,
Is there a way to optimize this type of queries? (column Value is
indexed):
SELECT Value FROM MyValues WHERE Value LIKE '%crashed.'
I've seen the document where 'begins with' queries can be optimized
using >= and < (end of the '4.0 The LIKE optimization' section):
6 matches
Mail list logo