Iterative search of a Teradata clob - for-loop

Question

We have an accountnumber stored in a Clob field in a table...we'll call it tbl_accountdetail. I need to pull back all records from tbl_accountdetail if the account numbers are in the results from another query...we'll call it sourcequery.
I can do this individually for each account number with:
Select * from Tbl_accountdetail where REGEXP_INSTR(CLOB,'accountnumber')>0
Naturally, my first thought was to do a cursor and loop through each account number from the sourcequery.
Declare #accountnumber varchar(30)
Declare Err_Cursor Cursor for
Select accountnumber from ErrorTable;
Open Err_Cursor;
Fetch next from Err_Cursor into #accountnumber;
While ##Fetch_status = 0
Begin
Select * from Tbl_accountdetail where REGEXP_INSTR(CLOB,#accountnumber)>0
Fetch next from Err_Cursor into #accountnumber
End;
Close Err_Cursor;
Deallocate Err_Cursor;
The more I read the more I'm confused about the best/most efficient way to get my desired results.
References to cursors all seem to require them to be included in a stored procedure and based on the simplicity, you wouldn't think this needs to be added to a sp. References to macros all seem to be macros that need to update/insert,etc. which I don't need. All I need to do is return the rows from Tbl_accountdetail that have the accountnumber somewhere in the clob.
I'm new to Teradata and Clobs. Can someone help me with the best way to search the clob? And to do so for a list of values?
Any help/suggestions greatly appreciated.

Answer 1

How is your CLOB data structured? Is the accountnumber field stored in a way that you can extract it using a searchable pattern -- i.e. accountnumber=<10-digit-#>?
If you want to search for multiple accountnumber values, then I think the best way is to extract the accountnumber value(s) from each CLOB and then search on them. Something like:
SELECT *
FROM Tbl_accountdetail
WHERE <extracted_accountnumber> IN (
SELECT account_number
FROM account_table
)
You are right that cursors are only used in stored procedures. The main purpose for them is to process each row of a result set individually and perform any additional logic, which you don't seem to need in your case. You could either put the SQL into a macro or just run it as-is.
Update
Assuming you only have one accountnum value in your CLOB field with a format of "accountnum": "123456789", you should be able to do something like this:
SELECT *
FROM Tbl_accountdetail
WHERE REGEXP_SUBSTR(myclob, '"accountnum":\s+"([0-9]+)"', 1, 1, NULL) IN (
SELECT account_number
FROM account_table
)
This should extract the first accountnumber match in your CLOB field and see if that value also exists in the IN sub-query.
I don't have a TD system to test on, so you may need to fiddle with the arguments a bit. Just replace myclob with the name of your CLOB field and update the sub-query in the IN(). Give that a try and let me know.
SQL Fiddle (Oracle)
Regexp Tester
Teradata - REGEXP_SUBSTR