如何从多列获取最小/最小值但使用mysql忽略0作为最小值/最小值？

Question

I am trying to fetch the result from 16 columns to find which column has the minimum value but that minimum value shouldn't be 0.

so I have tried my query something like below based on some suggestions but it doesn't work for me as it returns 0.

SELECT least(
    IFNULL(col_1,0),
    IFNULL(col_2,0),
    IFNULL(col_3,0),
    IFNULL(col_4,0),
    IFNULL(col_5,0),
    IFNULL(col_6,0),
    IFNULL(col_7,0),
    IFNULL(col_8,0),
    IFNULL(col_9,0),
    IFNULL(col_10,0),
    IFNULL(col_11,0),
    IFNULL(col_12,0),
    IFNULL(col_13,0),
    IFNULL(col_14,0),
    IFNULL(col_15,0),
    IFNULL(col_16,0)) as lowest
FROM `my_date_table` 
where date_id = '108'

Thanks and please don't mark as repeat question as I already go through some of the stack questions but couldn't find the right solution that I can apply on my query on php/mysql.

Answer 1

This is not an easy task. You cannot simply turn all 0 s to NULL s since LEAST() returns NULL if any of its arguments is NULL .

If you know in advance the maximum value accross all columns, you can use the trick commented by Barmar .

Else, assuming that you have a primary key called id in the table, a (suboptimal) solution solution could be to unpivot the table using UNION ALL , while turning 0 values to NULL values; then, you can take advantage of aggregate function MIN() , which ignores NULL values.

Example query for 5 columns:

SELECT id, MIN(col) lowest
FROM (
    SELECT id, NULLIF(col1, 0) col FROM mytable
    UNION ALL SELECT id, NULLIF(col2, 0) FROM mytable
    UNION ALL SELECT id, NULLIF(col3, 0) FROM mytable
    UNION ALL SELECT id, NULLIF(col4, 0) FROM mytable
    UNION ALL SELECT id, NULLIF(col5, 0) FROM mytable
) x
GROUP BY id

Demo on DB Fiddle

Sample data:

| id  | col1 | col2 | col3 | col4 | col5 |
| --- | ---- | ---- | ---- | ---- | ---- |
| 1   | 1    | 2    | 3    | 4    | 5    |
| 2   | 0    | 6    | 7    | 8    | 9    |
| 3   | 0    | 10   | 11   | 12   |      |

Results:

| id  | lowest |
| --- | ------ |
| 1   | 1      |
| 2   | 6      |
| 3   | 10     |

Answer 2

If there's a maximum possible value for the columns, you can replace any zero values with something above this maximum before calling LEAST . Then it will return the value of the least column that's lower than this maximum.

In the code below, I assume 999999 is large enough.

SELECT least(
    IF(IFNULL(col_1,0) = 0, 999999, col_1),
    IF(IFNULL(col_2,0) = 0, 999999, col_2),
    IF(IFNULL(col_3,0) = 0, 999999, col_3),
    IF(IFNULL(col_4,0) = 0, 999999, col_4),
    IF(IFNULL(col_5,0) = 0, 999999, col_5),
    IF(IFNULL(col_6,0) = 0, 999999, col_6),
    IF(IFNULL(col_7,0) = 0, 999999, col_7),
    IF(IFNULL(col_8,0) = 0, 999999, col_8),
    IF(IFNULL(col_9,0) = 0, 999999, col_9),
    IF(IFNULL(col_10,0) = 0, 999999, col_10),
    IF(IFNULL(col_11,0) = 0, 999999, col_11),
    IF(IFNULL(col_12,0) = 0, 999999, col_12),
    IF(IFNULL(col_13,0) = 0, 999999, col_13),
    IF(IFNULL(col_14,0) = 0, 999999, col_14),
    IF(IFNULL(col_15,0) = 0, 999999, col_15),
    IF(IFNULL(col_16,0) = 0, 999999, col_16)) as lowest
FROM `my_date_table` 
where date_id = '108'

Answer 3

One relatively simple (but expensive) method is to unpivot and aggregation:

select min(col), max(col)
from ((select col_1 as col from t) union all
      (select col_2 as col from t) union all
      . . .
      (select col_16 as col from t) 
     ) c
where col <> 0;

If you want to express this using least() , you can use coalesce() with nullif() :

select nullif( least( coalesce(nullif(col_1, 0), 999999),
                      coalesce(nullif(col_2, 0), 999999),
                      . . .
                    ), 999999
              )

This assumes that you have some reasonable maximum value to use for 999999 .