Sql-server – Execution plan flips Filter and Execute Scalar when using the PK, causes cast to fail

errorssql serversql-server-2005

I have some data tables that where imported in from a flat file source so everything is varchar(max) except FileRecordID which is a int that hold the row number the data came from in the flat file source (and is the primary key of the imported table.)

The following query works perfectly

SELECT *
  FROM [CLIENT]
  left outer join
  (
    select cast(ENTRY_CODE as int) as ENTRY_CODE, ENTRY_DESCRIPTION
    from [ENTRY]
    inner join [TABLE] on [ENTRY].ENTRY_TABLE = [Table].FileRecordID
    where ENTRY_RECD = 'A' and TABLE_RECD = 'A' and TABLE_CODE = 'CLTDISC' and ISNUMERIC(ENTRY_CODE) = 1
  ) as discounts on CLIENT_CLASS = ENTRY_CODE
  where Client_recd = 'A'
  --and client.[FileRecordID] = 10607

I get the exact number of rows I was expecting and everything is fine. However if I uncomment that last line so I only show the record from the client data table who has client.[FileRecordID] = 10607 it reports to me

Msg 245, Level 16, State 1, Line 1
Conversion failed when converting the varchar value 'BK/TN' to data type int.

The problem is from cast(ENTRY_CODE as int) as ENTRY_CODE in the first outer join. ENTRY can contain text in the ENTRY_CODE column, however all records that have ENTRY_TABLE set to the line number from TABLE where TABLE_CODE = 'CLTDISC' will always be numbers only. I perform the cast as ENTRY_CODE may or may not have leading 0's and spaces so I am trying to get them to be well formed.

What appears to me to be happening is adding client.[FileRecordID] = 10607 to the outer query causes the where clause not to be evaluated before the cast to int in the inner query.

I have tried things like adding and ISNUMERIC(ENTRY_CODE) = 1 to both the inner and outer queries but it has no affect.

How do I fix this?
Can anyone explain what is happening so I can understand how to avoid this in the future?

Non filtered query planed execution
enter image description here

Filtered (non working) query planned execution
enter image description here
Notice the reversed order of the Compute Scalar and the Filter

Other things I have tried that do not work:

Move TABLE_CODE = 'CLTDISC' in to the on clause in the inner join.
Added OPTION (FORCE ORDER) query hint to the query.

Update:
I found a solution but it is very "hackish" IMHO, I really would still like a explanation of how to avoid things like this in the future and how to do this correctly.

SELECT *
  FROM [CLIENT]
  left outer join
  (
    select ENTRY_CODE, ENTRY_DESCRIPTION
    from [ENTRY]
    inner join [TABLE] on [ENTRY].ENTRY_TABLE = [Table].FileRecordID
    where ENTRY_RECD = 'A' and TABLE_RECD = 'A' and TABLE_CODE = 'CLTDISC' and ISNUMERIC(ENTRY_CODE) = 1
  ) as discounts on CLIENT_CLASS = REPLACE(LTRIM(REPLACE(ENTRY_CODE, '0', ' ')), ' ', '0')
  where Client_recd = 'A'
  and client.[FileRecordID] = 10607

Best Answer

I'm sure there are dba.se and StackOverflow duplicates, but it was far faster for me to post this link:

https://feedback.azure.com/forums/908035-sql-server/suggestions/32912431-sql-server-should-not-raise-illogical-errors

In short, the Query Optimizer is really free to rewrite the plan as it sees best. Sometimes, it means a column is transformed (CAST -> int) close to retrieval so that it doesn't have to be carried many steps forward before performing the CAST.

The only sure-fire way of preventing the error is to look at the CAST expression itself, e.g.:

cast(CASE WHEN ISNUMERIC(ENTRY_CODE)=1 THEN ENTRY_CODE END as int)

Related Solutions

Sql-server – Using the name of a table contained in a column of another table

The only way to do this is to query SourceID in to a variable and then query the source table. At that point, you're better off pulling both into variablesm like this:

DECLARE @SourceTable NVARCHAR(255) --Note the length is arbitrary
,@SourceID NVARCHAR(255) --Assuming it's an INT, it doesn't matter
,@sql NVARCHAR(MAX)

SELECT @SourceTable = SourceID
,@SourceID = ReferenceID
FROM MyTable
WHERE SourceID = 'some_table'
AND ReferenceID = 1

SET @SQL = 'SELECT * FROM ' + @SourceTable + ' WHERE InstanceID = ' + @SourceID

EXEC(@sql)

You want the SourceID to be a string because otherwise you'd need to convert it to one anyway.

Sql-server – How (and why) does TOP impact an execution plan

I would have guessed that when a query includes TOP n the database engine would run the query ignoring the the TOP clause, and then at the end just shrink that result set down to the n number of rows that was requested. The graphical execution plan seems to indicate this is the case -- TOP is the "last" step. But it appears there is more going on.

The way the above is phrased makes me think you may have an incorrect mental picture of how a query executes. An operator in a query plan is not a step (where the full result set of a previous step is evaluated by the next one.

SQL Server uses a pipelined execution model, where each operator exposes methods like Init(), GetRow(), and Close(). As the GetRow() name suggests, an operator produces one row at a time on demand (as required by its parent operator). This is documented in the Books Online Logical and Physical Operators reference, with more detail in my blog post Why Query Plans Run Backwards. This row-at-a-time model is essential in forming a sound intuition for query execution.

My question is, how (and why) does a TOP n clause impact the execution plan of a query?

Some logical operations like TOP, semi joins and the FAST n query hint affect the way the query optimizer costs execution plan alternatives. The basic idea is that one possible plan shape might return the first n rows more quickly than a different plan that was optimized to return all rows.

For example, indexed nested loops join is often the fastest way to return a small number of rows, though hash or merge join with scans might be more efficient on larger sets. The way the query optimizer reasons about these choices is by setting a Row Goal at a particular point in the logical tree of operations.

A row goal modifies the way query plan alternatives are costed. The essence of it is that the optimizer starts by costing each operator as if the full result set were required, sets a row goal at the appropriate point, and then works back down the plan tree estimating the number of rows it expects to need to examine to meet the row goal.

For example, a logical TOP(10) sets a row goal of 10 at a particular point in the logical query tree. The costs of operators leading up to the row goal are modified to estimate how many rows they need to produce to meet the row goal. This calculation can become complex, so it is easier to understand all this with a fully worked example and annotated execution plans. Row goals can affect more than the choice of join type or whether seeks and lookups are preferred to scans. More details on that here.

As always, an execution plan selected on the basis of a row goal is subject to the optimizer's reasoning abilities and the quality of information provided to it. Not every plan with a row goal will produce the required number of rows faster in practice, but according to the costing model it will.

Where a row goal plan proves not to be faster, there are usually ways to modify the query or provide better information to the optimizer such that the naturally selected plan is best. Which option is appropriate in your case depends on the details of course. The row goal feature is generally very effective (though there is a bug to watch out for when used in parallel execution plans).

Your particular query and plan may not be suitable for detailed analysis here (by all means provide an actual execution plan if you wish) but hopefully the ideas outlined here will allow you to make forward progress.

Best Answer

Related Solutions

Sql-server – Using the name of a table contained in a column of another table

Sql-server – How (and why) does TOP impact an execution plan

Related Question