Modifiying and concatenating header titles chosen depending on body criteria - excel

Question

I have the following Excel table used to calculate the overtime hours worked.
I want the last two columns to be auto-generated, preferably with formulas, so that the user can see on which dates the employees have worked OT.
The example output for Bob would be:
Dates worked Normal OT : "2nd, 3rd, 4th, 5th, 10th & 12th"
Dates worked Double OT : "6th, 7th & 13th"
Please note that I don't have Excel 2016, and therefore can't use TEXTJOIN(). Also note that the dates for Week 1 and Week 2 are stored in number format, not date format, so the use of WEEKDAY() is also not possible.
P.S. I have already tried a TextJoin UDF but it doesn't seem to work since I have a lot of criteria within the formula.
The working formula for TEXTJOIN in Excel 2016 is this:
=TEXTJOIN(", ",TRUE,IF(WEEKDAY($B$3:$O$3,2)<6,IF($B5:$O5>0,TEXT($B$3:$O$3,"dd/mm/yyyy"),""),""))
This is using date formats. The UDF doesn't seem to work with these parameters.

Answer 1

I am ignoring everything in the PS at the end of your question as Jeeped is right in that it is contradictory to the main body and screenshot.
Without using a UDF, the formula would be much too complicated. Don't worry, I have supplied my own basic TEXTJOIN() UDF that definitely works.
All the following formulae need to be "array-entered" (by pressed Ctrl+Shift+Enter) in a single cell, and then copied/filled down. (Remember not to copy the starting { and ending }.)
The only difference in the two formulae is that one uses the comparison <5 whilst the other uses >=5, and each one refers to the appropriate "No. of Days" cell.
This first formula needs to be array-entered into cell T5 (Bob's "Dates Worked Normal OT") and then filled down:
{=SUBSTITUTE(TEXTJOIN(", ", TRUE, IF((MOD(COLUMN($B5:$O5)-COLUMN($B5), 7)<5)*($B5:$O5>0), $B$3:$O$3 & CHOOSE(IF($B$3:$O$3<4, $B$3:$O$3, IF($B$3:$O$3<21, 4, IF($B$3:$O$3<24, $B$3:$O$3-20, IF($B$3:$O$3<31, 4, 1)))), "st", "nd", "rd", "th"), "")), ", ", " & ", MAX(1, S5-1))}
The expanded, easier to read, version of the above formula (it will also work if you copy-paste it):
{=
SUBSTITUTE(
TEXTJOIN(
", ",
TRUE,
IF(
(MOD(COLUMN($B5:$O5)-COLUMN($B5), 7)<5)*($B5:$O5>0),
$B$3:$O$3
&
CHOOSE(
IF($B$3:$O$3<4, $B$3:$O$3, IF($B$3:$O$3<21, 4, IF($B$3:$O$3<24, $B$3:$O$3-20, IF($B$3:$O$3<31, 4, 1)))),
"st", "nd", "rd", "th"
),
""
)
),
", ",
" & ",
MAX(1, R5-1)
)}
This second formula needs to be array-entered into cell U5 (Bob's "Dates Worked Double OT") and then filled down:
{=SUBSTITUTE(TEXTJOIN(", ", TRUE, IF((MOD(COLUMN($B5:$O5)-COLUMN($B5), 7)>=5)*($B5:$O5>0), $B$3:$O$3 & CHOOSE(IF($B$3:$O$3<4, $B$3:$O$3, IF($B$3:$O$3<21, 4, IF($B$3:$O$3<24, $B$3:$O$3-20, IF($B$3:$O$3<31, 4, 1)))), "st", "nd", "rd", "th"), "")), ", ", " & ", MAX(1, S5-1))}
The expanded version of the above formula is:
{=
SUBSTITUTE(
TEXTJOIN(
", ",
TRUE,
IF(
(MOD(COLUMN($B5:$O5)-COLUMN($B5), 7)>=5)*($B5:$O5>0),
$B$3:$O$3
&
CHOOSE(
IF($B$3:$O$3<4, $B$3:$O$3, IF($B$3:$O$3<21, 4, IF($B$3:$O$3<24, $B$3:$O$3-20, IF($B$3:$O$3<31, 4, 1)))),
"st", "nd", "rd", "th"
),
""
)
),
", ",
" & ",
MAX(1, S5-1)
)}
Notes:
As mention above, the formulae rely on the availability/accuracy of the "No. of Days" cells in order to work correctly.
These formulae are pretty straightforward:
(MOD(COLUMN($B5:$O5)-COLUMN($B5), 7)>=5)*($B5:$O5>0) is just an array-formula friendly way of writing AND((…), (…));
The only sneaky thing is the use of the CHOOSE() function with four nested IF()s to select the ordinal indicator.
Don't forget to not include the { at the start, and the } at the end of the formulae when copy-pasting. These are just used to show that a formula needs to be array-entered.
My version of the TEXTJOIN UDF:
'============================================================================================
' Module : <any standard module>
' Version : 0.1.0
' Part : 1 of 1
' References : Optional - Microsoft VBScript Regular Expressions 5.5 [VBScript_RegExp_55]
' Source : https://stackoverflow.com/a/49218794/1961728
'============================================================================================
Public Function TEXTJOIN( _
ByRef delimiter As String, _
ByRef ignore_empty As Boolean, _
ByRef text1 As Variant _
) _
As String
Dim ƒ As Excel.WorksheetFunction: Set ƒ = Excel.WorksheetFunction
Const DELIMITER_ As String = "#"
Const PATTERN_ As String = "^(?:#)+|(?:#)+$|(#){2,}"
Static rexDelimiterEscaper As Object ' VBScript_RegExp_55.RegExp ' ## Object
Static rexEmptyIgnorer As Object ' VBScript_RegExp_55.RegExp ' ## Object
If rexEmptyIgnorer Is Nothing _
Then
Set rexEmptyIgnorer = CreateObject("VBScript.RegExp") ' New VBScript_RegExp_55.RegExp ' ## CreateObject("VBScript.RegExp")
With rexEmptyIgnorer
.Global = True
.Pattern = PATTERN_ ' Replacement = "$1"
End With
Set rexDelimiterEscaper = CreateObject("VBScript.RegExp") ' New VBScript_RegExp_55.RegExp ' ## CreateObject("VBScript.RegExp")
With rexDelimiterEscaper
.Global = True
.Pattern = "(.)" ' Replacement = "\$1"
End With
End If
Dim varText1 As Variant
Select Case TypeName(text1)
Case "Range":
varText1 = ƒ.Transpose(text1.Value2)
If text1.Rows.Count = 1 Then
varText1 = ƒ.Transpose(varText1)
If text1.Columns.Count = 1 Then varText1 = Array(varText1)
End If
Case "Variant()":
varText1 = text1
Case Else:
varText1 = Array(text1)
End Select
If ignore_empty _
Then
With rexEmptyIgnorer
.Pattern = Replace(PATTERN_, DELIMITER_, rexDelimiterEscaper.Replace(delimiter, "\$1"))
TEXTJOIN = .Replace(Join(varText1, delimiter), "$1")
End With
Else
TEXTJOIN = Join(varText1, delimiter)
End If
End Function
Notes:
This is not a proper poly-fill:
The first two arguments are not optional;
If you no not wish to use a delimiter, you must pass an empty string as the first parameter.
There is only one other (also required) argument allowed.
You can pass in anything for the third argument, except a multi-dimension array/range. Doing so will result in a #VALUE! error.
It should be very fast, especially for large inputs, as it doesn't use any loops. If you aren't ignoring empty values, it will be lightning fast. Ignoring them will be slower as a couple of regexes and an extra string manipulation have to be used as well.