Question

Suppose we have some data that looks like so

R_Id    Nm      Base       Dest    Proj      Cust_id  201203   201202   201201
MRBR    Bob     LONDON     UK      Project1  1        0        0        0
MRBU    Frank   LONDON     London  Project2  2        11.68    0        248.93
MRBU    Frank   LONDON     UK      Project3  1        7.4      4.8      0
MRGB    Barry   GUILDFORD  Hull    Project4  1        50.36    12.85    48.92
MRGB    Barry   GUILDFORD          Project5  1        0        177.31   0
MRGB    Barry   GUILDFORD  INTL    Project6  3        0        331.08   0

And suppose we have a lot more columns than above, but we’ve limited to a few for now.

I want to be able to use a where statment to only show rows where the row needs further investigation. This is done by saying “Where are there more than two large numbers next to each other in a row?” So I need to count the number of rows where the number is large.

The output should look like such, where I’ve explained what filtering I’m doing.

R_Id    Nm      Base       Dest    Proj      Cust_id  201203   201202   201201
MRBR    Bob     LONDON     UK      Project1  1        "Numbers not Large"
MRBU    Frank   LONDON     London  Project2  2        11.68    248.93   0   
MRBU    Frank   LONDON     UK      Project3  1        "Numbers not Large"
MRGB    Barry   GUILDFORD  Hull    Project4  1        50.36    12.85    48.92
MRGB    Barry   GUILDFORD          Project5  1        "Too few adjacent numbers"
MRGB    Barry   GUILDFORD  INTL    Project6  3        "Too few adjacent numbers"

It’s the case where there are too few adjacent numbers I’m trying to filter for. I need to count the numer of adjacent (or every other!) numbers in those specific columns.

I’ve looked at this question: Multiple Column Conditional Count SQL, but I don’t think I can use Count(*) as I get this error: You tried to execute a query thaty does not include the speicified expression ‘AT_RISK?’ as part of aggregate function. At risk is a column that just stores Yes/No and lives to the left of R_Id (not included above for brevity)

Can anyone help or at least point me in the right direction please? I’d really appreciate it. I’ve read the question above, and I’ve looked at how to use count in general, but this is really stumping me.

Answer 1

Well I can think of a somewhat ugly solution to this question, but it involves the use of a custom VBA function.

Concatenate and Test String

Your SQL statement should be something like:

SELECT * FROM tblName 
WHERE IsSeqHigh([201203] & ";" & [201202] & ";" & ..., 1000);

And then, in a VBA module we define:

Public Function IsSeqHigh(seq As String, thres As Double) As Boolean
    IsSeqHigh = False

    Dim valStrs() As String
    valStrs = Split(seq, ";")

    For n = 1 To UBound(valStrs) - 1
        If (valStrs(n) >= thres) And (valStrs(n + 1) >= thres) Then
            IsSeqHigh = True
            Exit For
        End If
    Next n
End Function

Another approach

Alternatively, if your schema is fixed and unlikely to change – and you have a primary key value, you can write a VBA function which takes the primary key value and scans the columns for the specific condition you are looking for.

In short, there is no good SQL-only solution that I can think of.