1

I have a query:

SELECT 
    count(session_id_open) as opens,
    count(session_id_visit) as visits,
    date_intervals_open,
    group_concat(date_intervals_visit)
FROM
    bla
GROUP BY date_intervals_open

I get following table. What I need is to find percentage wise the occurrences of each value appearing in group_concat. So, basically, I need to count number of values (date_intervals_visit) in each group (data_intervals_open)

opens   visits  date_intervals_open group_concat        
213    5        day (12-16)             evening (17-21),evening (17-21),day (12-16),day (12-16),day (12-16)
113    0        evening (17-21)         NULL
11     0        late evening (22-00)    NULL
396    12       morning (5-11)          morning (5-11),morning (5-11),morning (5-11),morning (5-11),morning (5-11),morning (5-11),morning (5-11),morning (5-11),morning (5-11),morning (5-11),morning (5-11),morning (5-11)
9      0        night (1-4)              NULL

That is approximately the table that I need to get. in First record evening has 40 because "evening (17-21)" appears twice and the number of all occurrences is 5. 2/5*100=40 

opens   visits  date_intervals_open evening(17-21)  day(12-16)  morning (5-11)  
213    5        day (12-16)             40          60        0
113    0        evening (17-21)         NULL        NULL      NULL
11     0        late evening (22-00)    NULL        NULL     NULL
396    12       morning (5-11)          0        0     100
9      0        night (1-4)             NULL

PS: I used group_concat just to visualize the value that I have there. I do not have to use it as it will be an additional effort to parse it afterwards.

Alina
  • 2,191
  • 3
  • 33
  • 68
  • Please provide `SHOW CREATE TABLE bla` and a few rows of sample data. It is as if you are depending on NULLs when doing `count(session_id_visit)`. This seems strange. – Rick James Dec 01 '16 at 00:25

3 Answers3

1

You essentially need a pivot, with some further calculations. The basis of my answer comes from the following excellent SO topic on pivoting records in MySQL. I assume that you have a fixed number of date_intervals_visit values, as these values seem to cover the entire day, therefore I use the conditional count approach with fixed number of counts. I'll add 2 categories in the sample code, you can extend it to cover all date_intervals_visit values.

SELECT 
    count(session_id_open) as opens,
    count(session_id_visit) as visits,
    date_intervals_open,
    round(sum(if(date_intervals_visit='morning (5-11)',1,0)) / count(session_id_visit) * 100,2) as `morning (5-11)`,
    round(sum(if(date_intervals_visit='day (12-16)',1,0)) / count(session_id_visit) * 100,2) as `day (12-16)`
FROM
    bla
GROUP BY date_intervals_open

If it is possible to have 0 visits by date_intervals_open value, then you need to check for 0 in the expression:

if(count(session_id_visit)=0, 0, <above formula>)
Community
  • 1
  • 1
Shadow
  • 33,525
  • 10
  • 51
  • 64
0
SELECT 
    count(session_id_open) as opens,
    @visits := count(session_id_visit) as visits,
    date_intervals_open,
    ROUND(100 * SUM(date_intervals_visit = 'evening(17-21)') / @visits) AS 'evening(17-21)',
    ROUND(100 * SUM(date_intervals_visit = 'day (12-16)') / @visits) AS 'day (12-16)',
    ROUND(100 * SUM(date_intervals_visit = 'morning (5-11)') / @visits)'morning (5-11)',
FROM
    bla
GROUP BY date_intervals_open
Rick James
  • 135,179
  • 13
  • 127
  • 222
-1

Use a function like this:

CREATE FUNCTION [dbo].[fn_SplitString](
    @InputStr   varchar(Max),
    @Seperator  varchar(10))
RETURNS @OutStrings TABLE (ItemNo int identity(1,1), Item varchar(256))

AS
BEGIN

    DECLARE @Str varchar(2000),
            @Poz int, @cnt int

    --DECLARE @OutStrings TABLE (Item varchar(2000))

    SELECT @Poz = CHARINDEX (@Seperator, @InputStr), @cnt = 0
    WHILE @Poz > 0 AND @cnt <= 10000
    BEGIN
        SELECT @Str = SubString(@InputStr, 1, @Poz - 1)
        INSERT INTO @OutStrings(Item) VALUES(@Str)

        SELECT @InputStr = Right(@Inputstr, Len(@InputStr) - (len(@Str) + len(@Seperator)))
        SELECT @Poz = CHARINDEX (@Seperator, @InputStr), @cnt = @cnt + 1
    END
    IF @InputStr <> ''
    BEGIN
        INSERT INTO @OutStrings(Item) VALUES(@InputStr)
    END

    RETURN
END

In the following way:

SELECT  opens,
        visits,
        date_intervals_open,
        [evening(17-21)]/[All]*100 AS [evening(17-21)],
        [day(12-16)]/[All]*100 AS [day(12-16)],
        [morning (5-11)]/[All]*100 AS [morning (5-11)]
FROM
    (   
    SELECT 
        count(session_id_open) as opens,
        count(session_id_visit) as visits,
        date_intervals_open,
        (SELECT COUNT(Item) FROM [dbo].[fn_SplitString](LTRIM(RTRIM(group_concat(date_intervals_visit))), ',') WHERE item  = 'evening(17-21)') AS [evening(17-21)],
        (SELECT COUNT(Item) FROM [dbo].[fn_SplitString](LTRIM(RTRIM(group_concat(date_intervals_visit))), ',') WHERE item = 'day(12-16)') AS [day(12-16)],
        (SELECT COUNT(Item) FROM [dbo].[fn_SplitString](LTRIM(RTRIM(group_concat(date_intervals_visit))), ',') WHERE item = 'morning (5-11)') AS [morning (5-11)],
        (SELECT COUNT(Item) FROM [dbo].[fn_SplitString](LTRIM(RTRIM(group_concat(date_intervals_visit))), ',')) AS [All]
    FROM
        bla
    GROUP BY date_intervals_open
    )blabla
Danie Schoeman
  • 180
  • 10