SUM,COUNT with UNION in MySql - mysql

Question

I have a table that has the following schema:
DATA | CAUSE_1 | TIME_1 | CAUSE_2 | TIME_2 | CAUSE_3 | TIME_3
The CAUSE.* field (VarChar) can not contain any string, and if so, the field TIME.* is 0.
I'm trying to create a query, but unfortunately without success, where I would have the result display in this form:
CAUSE | TOT_TIME | N_RIPET_CAUSE,
where:
In CAUSE I have a list of what is contained in CAUSE_1 ... CAUSE_3,
In TOT_TIME the sum of the values in TIME_1 ... TIME_3,
In N_RIPET_CAUSE the number of repetitions of each CAUSE.
I hope I explained.

Answer 1

try this
SELECT DATA ,CAUSE , TOT_TIME , N_RIPET_CAUSE
FROM ( select DATA, CONCAT(`CAUSE_1`,' ',`CAUSE_2`, ' ', `CAUSE_3`) as CAUSE ,
sum(`TIME_1` + `TIME_2` +`TIME_3`) as TOT_TIME ,
(count(`CAUSE_1`)+count(`CAUSE_2`)+count(`CAUSE_3`)) as N_RIPET_CAUSE
FROM your_table
group by DATA
) t
SEE SQLFIDDLE DEMO
EDIT.
try this
( select DATA , `CAUSE_1` as CAUSE ,
sum(`TIME_1` + `TIME_2` +`TIME_3`) as TOT_TIME ,
(count(`CAUSE_1`)+count(`CAUSE_2`)+count(`CAUSE_3`)) as N_RIPET_CAUSE
FROM Table1
group by DATA)
union all
(select DATA , `CAUSE_2` as CAUSE ,
sum(`TIME_1` + `TIME_2` +`TIME_3`) as TOT_TIME ,
(count(`CAUSE_1`)+count(`CAUSE_2`)+count(`CAUSE_3`)) as N_RIPET_CAUSE
FROM Table1
group by DATA )
union all
(select DATA , `CAUSE_3` as CAUSE ,
sum(`TIME_1` + `TIME_2` +`TIME_3`) as TOT_TIME ,
(count(`CAUSE_1`)+count(`CAUSE_2`)+count(`CAUSE_3`)) as N_RIPET_CAUSE
FROM Table1
group by DATA )
SQL DEMO HERE
EDIT:
try this due to your need
select cause, sum(time) Tot_time, count(cause) N_Ripet_Cause
from(
select cause_1 as cause, time_1 as time
from Table1
union all
select cause_2 as cause, time_2 as time
from Table1
union all
select cause_3 as cause, time_3 as time
from Table1
) t
group by cause
DEMO SQL FIDDLE

Answer 2

If you cannot change the table structure, then in order to get this result, you are going to need to unpivot the columns into rows.
MySQL does not have an unpivot function but this can be done using a UNION ALL query. You can then apply the aggregate to those values to get the final result:
select cause, sum(time) Tot_time, count(cause) N_Ripet_Cause
from
(
select data, cause_1 as cause, time_1 as time
from yourtable
union all
select data, cause_2 as cause, time_2 as time
from yourtable
union all
select data, cause_3 as cause, time_3 as time
from yourtable
) src
group by cause

Answer 3

You could make a select from union select like that:
select * from
(
select cause_1 as cause, time_1 as time from tableName
union
select cause_2 as cause, time_2 as time from tableName
union
select cause_3 as cause, time_3 as time from tableName
) as joinedValues
Then you could perform any actions from that select.
Like number of each clause:
select cause, count(cause) from
(
...
) as joinedValues
group by cause

Answer 4

Jack is on the mark - you've got too many possibly redundant cells in your table structure. Use relations to eliminate such occurrences.
DataTable
dID | Data
instancesTable
ID | dID | CAUSE | TIME
Then use NATURAL JOIN upon the two tables to extract the information;
SELECT * FROM DataTable NATURAL JOIN instancesTable WHERE dID=? LIMIT 3
This query will return a list of causes and times of whatever occurred on the ID of the 'Data' in the first table.
Edit: *N_RIPET_CAUSE* can be found using a SUM(CAUSE) on the dID.