Sql server conversion issues from Binary Data type to Numeric data


I am trying to persist values from one of the Sql server CDC table, which has column data type as "Binary(10)". I want to convert it into "Numeric" format and then convert it back to "Binary(10)".

For e.g.

declare @binary_data binary(10),@binary2 as binary(10)

select  @binary_data = 0x000000180003727C0006

Select Convert(int,0x00000018) as [PrecesionValue]
Select Convert(int,0x000) as [ScaleValue]

declare @Numeric as numeric(24,0) --Setting Numeric(PrecesionValue, ScaleValue)

Select @Numeric =Convert(numeric(24,0),@binary_data) 

Select @binary2 = Convert(binary(10),@Numeric) 

print @binary_data
print @Numeric
print @binary2


0x000000180003727C0006  //Initial binary data

393340                  //Converted Numeric value

0x0000180000017C000600  //Re-converted back to Binary value

If you see, the Re-convetred Binary value doesnt match the original value. Can you please check where I am going wrong?

asked on Stack Overflow Jun 28, 2018 by rajcool111 • edited Jun 28, 2018 by EzLo

1 Answer


To "translate" a 10-byte binary to something non-binary you'd need a data type, which has at least 10 bytes storage space. But you should not rely on the bit pattern used by SQL Server to store a complex data type.

I want to suggest two approaches:

--this is your binary
declare @binary1 binary(10) 
set @binary1 = 0x000000180003727C0006;
select  @binary1; 

--XML will translate binaries to base64 implicitly
declare @string varchar(100);
set @string = (SELECT @binary1 AS [*] FOR XML PATH(''),TYPE).value('.','varchar(100)');
select @string,LEN(@string);

--this is the approach to translate the base64 back to a binary
declare @binary2 binary(10);
set @binary2=CAST('<x>' + @string + '</x>' AS XML).value('.','varbinary(10)');
select @binary2; 

--the second approach is a GUID (16 byte)
declare @guid uniqueidentifier;
set @guid=@binary1
select @guid;

--a GUID is a simple chain of bytes, easy to cast
set @binary2=CAST(@guid AS VARBINARY(10));
select @binary2; 

UPDATE: one more idea came to my mind

There's a reason, why binary and string types are disussed in one topic. You can split your 10 byte binary in chunks and take them as separated numbers:

declare @binary1 binary(10) 
set @binary1 = 0x00000180003727C0006;
select  @binary1,SUBSTRING(@binary1,1,4) AS Byte0to3 
                ,SUBSTRING(@binary1,5,4) AS Byte4to8
                ,SUBSTRING(@binary1,10,2) AS Byte9to10;

declare @int1 INT, @int2 INT, @smallint smallint;
select  @int1 =     SUBSTRING(@binary1,1,4)  
       ,@int2 =     SUBSTRING(@binary1,5,4) 
       ,@smallint = SUBSTRING(@binary1,10,2);
select @int1,@int2,@smallint;

declare @binary2 binary(10);
set @binary2 = CAST(@int1 AS binary(4)) + CAST(@int2 AS binary(4)) + CAST(@smallint AS binary(2));
select @binary2;
answered on Stack Overflow Jul 1, 2018 by Shnugo • edited Jul 2, 2018 by Shnugo

User contributions licensed under CC BY-SA 3.0