SQL左外连接表

Question

I'm trying to pull out 3 columns from two tables, and I seem to be having a little problem.

SELECT a.[CreatedFromIp], b.[Token],  b.[MaskedP]
from [dbo].[Users] a LEFT OUTER JOIN [dbo].[Cards] b ON a.[Id]= b.[Id]
WHERE CreatedFromIp IS NOT NULL
Order by Token

but I keep getting Error converting data type nvarchar to bigint.

If I turn the tables around, I get all the token and masked info, but the IP result is just a whole list of NULL values.

Answer 1

You need to convert your both table id column to bigint - due to data type mismatch you are getting the error

SELECT a.[CreatedFromIp], b.[Token],  b.[MaskedP]
from [dbo].[Users] a LEFT OUTER JOIN [dbo].[Cards] b 
ON cast(a.[Id] as bigint)= cast(b.[Id] as bigint)
WHERE CreatedFromIp IS NOT NULL
Order by Token

Answer 2

SQL Server automatically converts away from strings when there is a comparison. The simple solution is to require that all columns that might be used in a join are of the same type. In this case, a simple foreign key definition would have helped when setting up the data model.

You can remove the error by converting to strings:

SELECT u.[CreatedFromIp], c.[Token], c.[MaskedP]
FROM [dbo].[Users] u LEFT OUTER JOIN
     [dbo].[Cards] c
      ON CONVERT(VARCHAR(255), u.[Id]) = CONVERT(VARCHAR(255), c.[Id])
WHERE CreatedFromIp IS NOT NULL
ORDER BY Token;

Of course, one of those type conversions is redundant, because the field is already a string.

Alternatively, you can convert the columns to integers -- using try_convert() or try_cast() :

SELECT u.[CreatedFromIp], c.[Token], c.[MaskedP]
FROM [dbo].[Users] u LEFT OUTER JOIN
     [dbo].[Cards] c
      ON TRY_CONVERT(int, u.[Id]) = TRY_CONVERT(int, c.[Id])
WHERE CreatedFromIp IS NOT NULL
ORDER BY Token;

Of course, the real solution is to fix the data model so the columns used in the JOIN s have the same type.