count difference on tables

[sds814]

On server A, when I do SELECT COUNT(*) FROM tblzCdCustomFieldLabels I get 574,228. On server B, when I do SELECT COUNT(*) FROM tblzCdCustomFieldLabels I get 2,533,100. So the difference being 1,958,872. However, when I do

SELECT *
FROM [A].databasename.dbo.tblzCdCustomFieldLabels prod
WHERE NOT EXISTS (SELECT * 				  FROM [B].databasename.dbo.tblzCdCustomFieldLabels dev
				  WHERE dev.eff_dt = prod.eff_dt AND dev.prod_ID = prod.prod_ID 
					AND dev.prodname_CD = prod.prodname_CD AND dev.CO_CD = prod.CO_CD)

I get the difference as 1,985,332. The 4 fields make up the primary key for the table tblzCdCustomFieldLabels. Shouldn't the difference be the same? Is the difference due to having more null values in the rows on server B?

 

[jbenson001]

If there are NULLs in any of the columns you are joining on, then the comparisons will evaluate to false and not give you the totals you expect

 

[sds814]

The columns that I'm joining on combine as the primary key so they couldn't have Null values.

 

[haroldholt]

Perhaps there are 26460 rows in B that do not exist in A ?

 

[RyanEK]

Your query determines records that do not exist on another server, not the difference in totals.