Row number per group in mysql

I want to produce query result base on this scenario that can create row number according to crew_id and type.

id   crew_id   amount    type
 1      4       1000      AUB
 2      4       1500      AUB
 3      5       8000      CA
 4      4       1000      CA
 5      5       1000      AUB
 6      6       3000      AUB
 7      4       2000      CA
 8      6       3500      AUB
 9      4       5000      AUB
 10     5       9000      CA
 11     5       1000      CA

OUTPUT must be the ff:

id    crew_id   amount   type    row_number
 1      4       1000      AUB        1    
 2      4       1500      AUB        2 
 9      4       5000      AUB        3
 4      4       1000      CA         1
 7      4       2000      CA         2
 5      5       1000      AUB        1
 3      5       8000      CA         1
 10     5       9000      CA         2
 11     5       1000      CA         3
 6      6       3000      AUB        1
 6      6       3000      AUB        2

I want a single select statement only in this output


Please go through my fiddle

This One Last Tried

    SELECT    id,
              crew_id,
              amount,
              type,
             ( 
                CASE type 
                WHEN @curType 
                THEN @curRow := @curRow + 1 
                ELSE @curRow := 1 AND @curType := type END
              ) + 1 AS rank
    FROM      Table1 p,
              (SELECT @curRow := 0, @curType := '') r
   ORDER BY  crew_id,type asc;

The question is quite old. But I would like to post it in case someone will have a same problem.

First of all, described answers do not work correct. For example, for

id   crew_id   amount    type
1      4       1000      AUB
2      4       1500      AUB
5      5       1000      AUB
6      6       3000      AUB
8      6       3500      AUB
9      4       5000      AUB

(I just removed rows with type 'CA') the result table will be

id   crew_id   amount    rank   type
1      4       1000      1      AUB
2      4       1500      2      AUB
9      4       5000      3      AUB
5      5       1000      4      AUB
6      6       3000      5      AUB
8      6       3500      6      AUB

So in fact it doesn't use both crew_id and type, it just uses type.

Here is how I solved this problem (probably there is a more elegant way to do it than use two nested 'CASE's, but you get the idea):

SELECT id,
    amount,
    CASE crew_id 
        WHEN @curCrewId THEN
            CASE type 
                WHEN @curType THEN @curRow := @curRow + 1 
                ELSE @curRow := 1
            END
        ELSE @curRow :=1
    END AS rank,
    @curCrewId := crew_id AS crew_id,
    @curType := type AS type
FROM Table1 p
JOIN (SELECT @curRow := 0, @curCrewId := 0, @curType := '') r
ORDER BY crew_id, type

The main idea remain. I just added a variable @curCrewId. If someone need to use 3 variables for grouping, so just use 3 variables and 3 nested 'CASE's. :)


@Janty's answer does not work when the type starts with a number (the rank starts at 2 instead of 1).

Use the following instead:

SELECT id,
     crew_id,
     amount,
     CASE type 
         WHEN @curType THEN @curRow := @curRow + 1 
         ELSE @curRow := 1
     END AS rank,
     @curType := type AS type
FROM Table1 p
JOIN (SELECT @curRow := 0, @curType := '') r
ORDER BY crew_id, type

In addition to the answer of @Janty here is a solution if you want to UDATE your table with the rownumber:

UPDATE myTable mt,(SELECT @curRow := 0, @curType := '') r SET type=
    ( 
        CASE type 
            WHEN @curType
        THEN @curRow := @curRow + 1 
        ELSE @curRow := 1 AND @curType := type END
      )
;

As janty too crewId is not within. Use a second "case" for that as mentioned in the other answers.