Question

I have a column called image_small for a table images which will have information like this:

image_small
/images/menu_items/148/5a13140f8ef4503f024d06231037c69e_medium.jpg
/images/menu_items/152/5a14444444444444424d06231037c69e_medium.jpg
 ...

I’d like to select out a the count of unique values for the menu item value (148) above (assume 148-400) on Mysql 5.1 using innodb so that output would be like:

val    count  
148    5  
101    4  
152    1  

The integer values (148,101,152) don’t exist in a column and would rather not loop through.

Any ideas on how to do this? Would seem like would need to create a notion of a variable for the ‘148’ and then a group by but a little clueless.

thx

Answer 1

I’m supposing that all your paths have the same format.

 select 
 substring_index(substring_index(str,'/',4),'/',-1) as val,
 count(*)  as `count`
 from table
 group by val

It seems to work but I’m afraid about performance on large tables. Do some tests.
Could you add a field with this value? I think it would be better.