Skip to main content
Announcements
NEW: Seamless Public Data Sharing with Qlik's New Anonymous Access Capability: TELL ME MORE!
cancel
Showing results for 
Search instead for 
Did you mean: 
Sloane
Contributor II
Contributor II

Dynamically Splitting fields

Hello everyone, 

 

I am quite new to qlik and would appreciate if anyone has any useful tip on how to achieve dynamically splitting columns into multiple columns based on delimeters.

 

here is a sample of what I have:

Sloane_0-1679923058438.png

and I am trying to achieve this:

Sloane_1-1679923114482.png

Any help would be greatly appreciated. 

Labels (4)
1 Solution

Accepted Solutions
MartW
Partner - Specialist
Partner - Specialist

@Sloane ,

 

can you try this script?

Table:
Load 
	*,
	len(KeepChar(Score,','))		as Check
;
LOAD
    ID,
    Score,
    Subject
FROM [lib://tmp/tmp.xlsx]
(ooxml, embedded labels, table is Sheet1);


Check:
NoConcatenate
Load max(Check) as max Resident Table;
drop field Check from Table;

let vMax = peek('max',0,'Check')+1;
Trace >>>$(vMax);

for vCounter = 1 to $(vMax)
	left join(Table)
    LOAD
        ID,
        subfield(Score,',',$(vCounter)) as Score$(vCounter),
        subfield(Subject,',',$(vCounter)) as Subject$(vCounter)
    FROM [lib://tmp/tmp.xlsx]
    (ooxml, embedded labels, table is Sheet1);
next

drop table Check;
let vMax = null();
let vCounter = null();

 

I've tested this with the data you showed in the screenshot and that worked.

 

MartW_0-1679927849269.png

 

View solution in original post

6 Replies
marcus_sommer

You could use subfield() with just two parameters to split the values into separate records. To create from them separate fields you may use a generic load - but within many scenarios it's better not to create such artificial crosstables else just keeping the information within a single field and using a pivot within the UI to show the horizontally information.

Sloane
Contributor II
Contributor II
Author

Thanks Marcus, I tried the subfield function and that split into rows which is not entirely bad but it gets tricky when I need the matching split scores for each subject. 

marcus_sommer

Maybe it's better to control the subfield() explicitly which makes it easier to split multiple fields at the same time and provides also a counter. This might be done with something like this:

load subfield(Score & ',' , ',', iterno()) as Score, iterno() as IterNo
from Source while iterno() <= substringcount(Score & ',', ',');

Sloane
Contributor II
Contributor II
Author

Thank you Marcus.

 

Surprisingly, I have used this exact method and it sadly didnt work. From your suggestion though, I see that you are splitting only the Score but I would ideally like the subjects to be split at the same time

MartW
Partner - Specialist
Partner - Specialist

@Sloane ,

 

can you try this script?

Table:
Load 
	*,
	len(KeepChar(Score,','))		as Check
;
LOAD
    ID,
    Score,
    Subject
FROM [lib://tmp/tmp.xlsx]
(ooxml, embedded labels, table is Sheet1);


Check:
NoConcatenate
Load max(Check) as max Resident Table;
drop field Check from Table;

let vMax = peek('max',0,'Check')+1;
Trace >>>$(vMax);

for vCounter = 1 to $(vMax)
	left join(Table)
    LOAD
        ID,
        subfield(Score,',',$(vCounter)) as Score$(vCounter),
        subfield(Subject,',',$(vCounter)) as Subject$(vCounter)
    FROM [lib://tmp/tmp.xlsx]
    (ooxml, embedded labels, table is Sheet1);
next

drop table Check;
let vMax = null();
let vCounter = null();

 

I've tested this with the data you showed in the screenshot and that worked.

 

MartW_0-1679927849269.png

 

Sloane
Contributor II
Contributor II
Author

Thanks a lot @MartW !!

 

this definitely helped a lot!