I just connect to SQL Server and invoke the following command:

select * from sys.sysrscols

The result table contain a field hbcolid, which seem to be identical to rscolid. I think rscolid is surely for rowset column id. Then what is the meaning and usage of hbcolid?

Is there any document on this?


Given that this is an internal system table that you can only access when connecting via the DAC, no, there is no documentation on this. What are you using this table for?

I found several instances here where rscolid was not identical to hbcolid. For example:

SELECT rsid, rscolid, hbcolid
  FROM sys.sysrscols
  -- WHERE rscolid <> hbcolid
  WHERE rsid = 72057594038714368;

rsid                 rscolid   hbcolid
-----------------    -------   -------
72057594038714368    4         2 

Where does this row come from?

  FROM sys.partitions 
  WHERE partition_id = 72057594038714368;


So what indexes might be involved?

SELECT index_id, name
  FROM sys.indexes 
  WHERE object_id = OBJECT_ID('sys.queue_messages_1003150619');

index_id   name
--------   ---------------------
1          queue_clustered_index
2          queue_secondary_index

So going on a guess, I ran this query:

  ic1.column_id, -- column id in the table
  ic1.key_ordinal -- column order in the index
FROM sys.all_columns AS c
INNER JOIN sys.index_columns AS ic1
ON c.[object_id] = ic1.object_id
AND c.column_id = ic1.column_id
AND ic1.column_id <> ic1.key_ordinal
WHERE c.object_id = OBJECT_ID('sys.queue_messages_1003150619');

One of the rows here:

index_id   name                   column_id   key_ordinal
--------   ---------------------  ---------   -----------
1          conversation_group_id  4           2

So, without doing much more investigation, my initial guess is that rscolid is the column id within the base table, and hbcolid is the ordinal position of that column in the index. This holds true for 5 of the columns in the clustered index of this queue table - they all have identical mismatches in sys.sysrscols, and also for one of the columns in the non-clustered index.

But again, whether I'm right or wrong, I'm not sure what you could possibly do with that information: what are you using this table for?


Please refer to below links:

dbForums TechNet

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.