Comma Separated Value Taking too Much Time to Execute

Question

Hi,
select ES_DIAGNOSIS_CODE DC from ecg_study WHERE PS_PROTOCOL_ID LIKE 'H6L-MC-LFAN'
The above query returns comma separated value from the above query.
I am using the query below to split the comma separated value but the below query is taking lot of time to return the data.
SELECT
select DC from (
with t as ( select ES_DIAGNOSIS_CODE DC from ecg_study WHERE PS_PROTOCOL_ID LIKE 'H6L-MC-LFAN' )
select REGEXP_SUBSTR (DC, '[^,]+', 1, level) DC from t
connect by level <= length(regexp_replace(DC,'[^,]*'))+1 )
Please suggest me is there any alternative way to do this comma separated value.
Thanks
Sudhir

Answer 1

Nikolay Savvinov wrote:
Hi BluShadow,
I know that this function is fast with varchar2 strings from several years of using it. With CLOBs one may need something faster, but the OP didn't menion CLOBs.
Best regards,
NikolayJust because you perceive it to be fast doesn't mean it's faster than doing it in SQL alone.
For starters you are context switching from the SQL engine to PL/SQL to call it.
Then in your code you are doing this...
select substr(v_str,v_last_break+1, decode(v_nxt_break,0,v_length, v_nxt_break-v_last_break-1)) into v_result from dual;which is context switching back from the PL/SQL engine to the SQL engine for each entry in the string.
Why people do that I don't know... when PL/SQL alone could do it without a context switch e.g.
v_result := substr(v_str,v_last_break+1, case when v_nxt_break = 0 then v_length else v_nxt_break-v_last_break-1 end);So, if you still think it's faster than pure SQL (which is what the OP is using), please go ahead and prove it to us.