How to group 3 columns into a single one in SQL - sql

Question

I have a table that has three columns: myColumn, myColumn2 and myColumn3. The three columns can have a different value but they are all the same type.
I need to do a join such that I merge all the items in the three columns and then group by
If I had a single column, it would be done like this:
select distinct (myColumn), COUNT(myColumn) as 'TotalF'
from FormularioCorto
where idEvento = #idEvento and myColumn <> ''
group by myColumn
The question is:
How do I do a TSQL statement that will allow for me to join the three columns and then group by?
Edit: here is the sample data
I have a table that has 3 columns with the same type
columna columnb columnc
a a b
b a b
c c d
I need to run a tsql that will merge the 3 columns and group by, to get a result like
newcolumn count
a 3
b 3
c 2
d 1
Any ideas?
Thanks in advance

Answer 1

It is hard to tell without some sample data and desired output but a guess would be something like:
SELECT MyCol as MyColumn, Count(*) as 'TotalF'
FROM
(
SELECT myColumn MyCol from FormularioCorto
WHERE idEvento = #idEvento AND myColumn <> ''
UNION ALL
SELECT myColumn2 MyCol from FormularioCorto
WHERE idEvento = #idEvento AND myColumn2 <> ''
UNION ALL
SELECT myColumn3 MyCol from FormularioCorto
WHERE idEvento = #idEvento) AND myColumn3 <> ''
) A
GROUP BY MyCol;
Or for just the example data and result:
SELECT NewColumn, Count(*) as Count FROM
(
SELECT columna as NewColumn FROM SomeTable
UNION ALL
SELECT columnb as NewColumn FROM SomeTable
UNION ALL
SELECT columnc as NewColumn FROM SomeTable
) A
Group by NewColumn

Answer 2

I think you mean concatenate rather than JOIN. JOIN has a very specific meaning with SQL. Use the + operator to concatenate string fields.

Answer 3

CONCAT(`myColumn`, `myColumn2`, `myColumn3`) as joinedField
The CONCAT() function concatenates fields and strings.
Info:
http://dev.mysql.com/doc/refman/5.0/en/string-functions.html#function_concat
Note: You should reconsider your structure though, as this will prove to be painfully slow as you add more records.

Answer 4

select cast(myColumn as nvarchar) + CAST(COUNT(myColumn)as nvarchar) as 'TotalF'
from FormularioCorto
where idEvento = #idEvento and myColumn <> ''
group by myColumn

Answer 5

If I am understanding your question right, then you simply want 2 columns (one identifying the id you are counting and one giving the count for the identifier. If this is correct, simply use your existing query for each column and union the results.
For example;
SELECT myColumn AS Id, COUNT(myColumn) AS IdCount
FROM FormularioCorto
where idEvento = #idEvento and myColumn <> ''
GROUP BY myColumn
UNION
SELECT myColumn2 AS Id, COUNT(myColumn2) AS IdCount
FROM FormularioCorto
where idEvento = #idEvento and myColumn2 <> ''
GROUP BY myColumn2
etc etc.

Answer 6

Try this:
SELECT myColumn1 + myColumn2 + myColumn3 AS AllThreeFields
FROM FormularioCorto
WHERE 1=1
GROUP BY myColumn1 + myColumn2 + myColumn3
It won't be fast but it should work. If you need to delimit the data in the columns with other characters you can try this:
SELECT myColumn1 + ' ' + myColumn2 + ' ' + myColumn3 AS AllThreeFields
FROM FormularioCorto
WHERE 1=1
GROUP BY myColumn1 + myColumn2 + myColumn3
If the column types are not the same (you said they were), you will need to cast them to a similar type.