将列值转换为行

Question

uid   pid   mail   value
 1     78    27     Nairobi
 2     78    27     Milimani
 3     78    27     Criminal
 4     78    27     1427932800

I have a DB table above and only need the 'value' column values. I want to have the column values display in rows (not comma separated) for a cross-tab report. my ideal result would be:

        **Nairobi  Milimani  Criminal  1427932800**

The matching 'pid' and 'mail' means that the corresponding 'value' is from a single submission and a change in pid and mail (not captured here) is a new submission!

so how do I write an sql for converting the 'value' column values to row values? any help much appreciated.

'Pivot' has not really helped or i'm probably doing it wrongly.!!

Answer 1

In SQL Server, you could easily pivot your columns something like this:

select 
     pvt.[Nairobi],
     pvt.[Milimani],
     pvt.[Criminal],
     pvt.[1427932800] 
from 
(select * from test) t
PIVOT
(
count(uid) for [value] in ([Nairobi],[Milimani],[Criminal],[1427932800]) 
        --change the count function and column to match your needs
) as pvt;

Also, as you may see above, you would need to use some kind of aggregate function to use pivot. Hope this helps!

SQL Fiddle Demo

UPDATE

After re-reading your question, I figured that you might be instead looking for something like this:

SELECT 
     ltrim(substring(t.concat_values, 1, charindex(' ', t.concat_values, 1))) AS first_col
    ,ltrim(substring(t.concat_values, dbo.GetNthCharacterOccurrence(t.concat_values, ' ', 1), (dbo.GetNthCharacterOccurrence(t.concat_values, ' ', 2) - dbo.GetNthCharacterOccurrence(t.concat_values, ' ', 1)))) AS second_col
    ,ltrim(substring(t.concat_values, dbo.GetNthCharacterOccurrence(t.concat_values, ' ', 2), (dbo.GetNthCharacterOccurrence(t.concat_values, ' ', 3) - dbo.GetNthCharacterOccurrence(t.concat_values, ' ', 2)))) AS third_col
    ,ltrim(substring(t.concat_values, dbo.GetNthCharacterOccurrence(t.concat_values, ' ', 3), len(t.concat_values) - dbo.GetNthCharacterOccurrence(t.concat_values, ' ', 3)+1)) AS fourth_col
FROM (
    SELECT DISTINCT stuff((
                SELECT ' ' + t2.[value]
                FROM test t2
                WHERE t1.pid = t2.pid
                    AND t1.mail = t2.mail
                FOR XML path('')
                ), 1, 1, '') AS concat_values
    FROM test t1
    ) t;

SQL Fiddle Demo 2

The trick for this method is to initially create a comma separated list of values using the STUFF function with XML Path . Then, break the string based on the position of the string separator which in this case I used space (' '). To find the nth occurrence of space, I "borrowed" a function that was written by Tavis Lovell originally in a blog. To split the values into multiple columns, I used substring , charindex and the user defined function above as needed.

Answer 2

MySQL Version

Using group_concat and substring_index functions, you could easily convert your string values into multiple columns like this:

SELECT 
     SUBSTRING_INDEX(SUBSTRING_INDEX(concat_values, ',',1), ',', -1) as first_col,
     SUBSTRING_INDEX(SUBSTRING_INDEX(concat_values, ',',2), ',', -1) as second_col,
     SUBSTRING_INDEX(SUBSTRING_INDEX(concat_values, ',',3), ',', -1) as third_col,
     SUBSTRING_INDEX(SUBSTRING_INDEX(concat_values, ',',4), ',', -1) as fourth_col

FROM (
     SELECT GROUP_CONCAT(value) as concat_values
    FROM test t1  --change this to match your table name
    ) t;

MySQL Demo