Сомнение возникло в рекомендации "нужно использовать collation, который позволяет хранить в юникоде (например, Cyrilic_General_CI_AS)"

[VitaliyK]

вот еще планы из примеров выше

~300ms

Цитата:

select
*
from SalesPickingListJournalLine isum (nolock)
where PickingListId like '%01Н0002320%'
select db_name(),getdate(), @d, datediff(ms,@d,getdate())
go

Цитата:

|--Parallelism(Gather Streams)
|--Nested Loops(Inner Join, OUTER REFERENCES-([Bmk1000]))
|--Index Scan(OBJECT-([db_cyr].[dbo].[SALESPICKINGLISTJOURNALLINE].[I_806DELIVERYREQUESTIDX] AS [isum]), WHERE-([db_cyr].[dbo].[SALESPICKINGLISTJOURNALLINE].[PICKINGLISTID] as [isum].[PICKINGLISTID] like '%01Н0002320%'))
|--RID Lookup(OBJECT-([db_cyr].[dbo].[SALESPICKINGLISTJOURNALLINE] AS [isum]), SEEK-([Bmk1000]=[Bmk1000]) LOOKUP ORDERED FORWARD)

~30-40ms:

Цитата:

select
*
from SalesPickingListJournalLine isum (nolock)
where TST_01 like '%01Н0002320%'

Цитата:

|--Compute Scalar(DEFINE-([isum].[TST_01]=[db_cyr].[dbo].[SALESPICKINGLISTJOURNALLINE].[TST_01] as [isum].[TST_01]))
|--Parallelism(Gather Streams)
|--Nested Loops(Inner Join, OUTER REFERENCES-([Bmk1000]))
|--Compute Scalar(DEFINE-([isum].[TST_01]=CONVERT(varchar(100),[db_cyr].[dbo].[SALESPICKINGLISTJOURNALLINE].[PICKINGLISTID] as [isum].[PICKINGLISTID],0)))
| |--Index Scan(OBJECT-([db_cyr].[dbo].[SALESPICKINGLISTJOURNALLINE].[I_806DELIVERYREQUESTIDX] AS [isum]), WHERE-(CONVERT(varchar(100),[db_cyr].[dbo].[SALESPICKINGLISTJOURNALLINE].[PICKINGLISTID] as [isum].[PICKINGLISTID],0) like '%01Н0002320%'))
|--RID Lookup(OBJECT-([db_cyr].[dbo].[SALESPICKINGLISTJOURNALLINE] AS [isum]), SEEK-([Bmk1000]=[Bmk1000]) LOOKUP ORDERED FORWARD)

TST_01:
[TST_01] AS ([PickingListId] collate SQL_Latin1_General_CP1251_CI_AS)