Regexp_substr and connect by level.

Question

Hi
I have a a below query and out put also given below
Query:
select distinct cont_id,ct.TAG245 from contact_tag ct where ct.CONT_ID='AEGA-3R43UM'
OUTPUT
Cont_id TAG245
AEGA-3R43UM     060|02.00|060|02.00|060|02.00|060|02.00
In the above output for TAG245 we have two values first value is percentage and second value is frequency.
I need to make all frequency values sum.(ie 02.00+02.00+02.00+02.00)
Expected output
Plannedcall
6
Step 1
First I am approaching the above output dividing the each pipe symbols into each row using the below query but it is not working it is taking long time.Can you advice me on this so that i can goto next step:2.
The query
SELECT regexp_substr (ct.TAG245, '[|,]+', 1, level) frequency
FROM contact_tag ct
--where ct.CONT_ID='AEGA-3R43UM'
CONNECT BY LEVEL <= LENGTH (REGEXP_REPLACE (ct.TAG245, '[|,]+')) + 1
order by 1;
Step: 2
Once i get the output I will sum the even rows because all evenrows has frequency values.
Thanks

Answer 1

876639 wrote:
Hi Blu Shadow
When I Run the below query
SELECT level as lvl, regexp_substr (ct.TAG245, '[^|]+', 1, level) frequency
FROM contact_tag ct
WHERE cont_id = 'AEGA-3R43UM'
CONNECT BY regexp_substr(ct.TAG245, '[^|]+', 1, level) is not null
it is taking long time, because you have used the with clause, here i am getting the values directly from Table Contact _tag,
could you telll me? why it is like that.Do you have an index on cont_id or is it the primary key?
How much data is in the table?
We can't tell you why it's going slow unless you provide us with sufficient information.
See the threads linked to by the following post:
{message:id=9360003}