Question

It’s my table rows:

++ id ---- text ++++++++++++++
-- 1  ---- '90','80,'50' -----
-- 2  ---- '30','2','1_2' --
-- 3  ---- '10_2','5_3' -----

as you see, text contains 2 types of numbers, one doesn’t have underscore, and the other does.

I want to select rows which have at least one number without underscore (type 1). Something like this: (result-set)

++ id ---- text ++++++++++++++
-- 1  ---- '90','80,'50' -----
-- 2  ---- '30','2','1_2' --

(3 is ignored)

How to do that? (I think it’s possible with NOT LIKE, but I don’t know how to write)

Answer 1

The below query counts the number of commas in the strings, number of distinct numbers can be calculated as 1 more than the number of commas as the numbers are separated by commas, number of underscores in the string:

select id,
len(text) - len(replace(text,',','')) as count_of_commas,
len(text) - len(replace(text,',','')) + 1 as count_of_number,
len(text) - len(replace(text,'_','')) as count_of_underscore,
len(text) - len(replace(text,',','')) + 1 - (len(text) - len(replace(text,'_',''))) as zero_if_no_number_without_underscore_exists
from t1

The above query gives the following results:

Now using the logic of above query, following query can be used to get the desired result:

select * from t1
where len(text) - len(replace(text,',','')) + 1 - (len(text) - len(replace(text,'_',''))) != 0

i.e. it returns the rows where atleast one number exists without the underscore.