Skip to main content
cancel
Showing results for 
Search instead for 
Did you mean: 
takama13
Contributor II
Contributor II

Intervalmatch to correct values

Hi all,

My database is a set of 2 Excel files.

Some fields are the same (and used to link the 2 tables), some fields are different.

In each table, I have a field with a date of events.

But some data (Service Code) are wrong and I would like to correct them based on an Excel file I created with range dates.

Here below are my tables :

Table1 :  

Service CodeDateEventShould be
MAN01/06/2016Event1MAGE
MAN01/05/2018Event2MTS
COL01/06/2016Event3COLU
COL01/06/2018Event4CJS
VES01/01/2016Event5VESP
VES01/07/2018Event6VESP
MTS01/06/2016Event7MTS
MTS01/06/2018Event8CTPS

Table2 :  

Service CodeDateEventBShould be
MAN01/06/2016Event1MAGE
MAN01/05/2018Event2MTS
COL01/06/2016Event3COLU
COL01/06/2018Event4CJS
VES01/01/2016Event5VESP
VES01/07/2018Event6VESP
MTS01/06/2016Event7MTS
MTS01/06/2018Event8CTPS
VIC01/02/2017Event9VIC

Service Code correction:    

Service CodeFromToNew Service Code
MAN01/01/201501/04/2017MAGE
MAN02/04/201705/05/2018MTS
COL01/01/201504/04/2017COLU
COL05/04/201701/01/2019CJS
VES01/01/201501/01/2019VESP
MTS26/04/201801/01/2019CTPS

Here is the code I use to try and correct the tables :

Table1:

LOAD

     [Service Code],

     Date,

     Event,

     [Should be]

FROM

(ooxml, embedded labels, table is Feuil1);

Table2:

LOAD

     [Service Code],

     Date,

     EventB,

     [Should be] as [Should be 2]

FROM

(ooxml, embedded labels, table is Feuil1);

'Service Code correction':

LOAD

     [Service Code],

     From,

     To,

     [New Service Code]

FROM

(ooxml, embedded labels, table is Feuil1);

Inner Join ('Table1')

IntervalMatch ([Date],[Service Code])

LOAD [From], [To], [Service Code] Resident 'Service Code correction';

Right Join('Table1')

LOAD * Resident 'Service Code correction';

Inner Join ('Table2')

IntervalMatch ([Date],[Service Code])

LOAD [From], [To], [Service Code] Resident 'Service Code correction';

Left Join('Table2')

LOAD * Resident 'Service Code correction';

DROP Table 'Service Code correction';

DROP Field [From];

DROP Field [To];

The result is correct unless the fact that the lines with Service Code not present in 'Service Code correction' are removed.

I would like to keep them.

What am I doing wrong ?

Result for Table2:

     

DateEventBService CodeShould be 2New Service Code
01/01/2016Event5VESVESPVESP
01/06/2016Event1MANMAGEMAGE
01/06/2016Event3COLCOLUCOLU
01/05/2018Event2MANMTSMTS
01/06/2018Event4COLCJSCJS
01/06/2018Event8MTSCTPSCTPS
01/07/2018Event6VESVESPVESP

We can notice that the line with Service Code "VIC" disappeared. I would like to keep this line with 'New Service Code' column empty.

Thanks

1 Solution

Accepted Solutions
Gysbert_Wassenaar

Inner Join ('Table2')

IntervalMatch ([Date],[Service Code])

LOAD [From], [To], [Service Code] Resident 'Service Code correction';

Left Join('Table2')

LOAD * Resident 'Service Code correction';

Change the Inner Join to a Left Join if you want to keep all the records from Table2


talk is cheap, supply exceeds demand

View solution in original post

12 Replies
Gysbert_Wassenaar

Inner Join ('Table2')

IntervalMatch ([Date],[Service Code])

LOAD [From], [To], [Service Code] Resident 'Service Code correction';

Left Join('Table2')

LOAD * Resident 'Service Code correction';

Change the Inner Join to a Left Join if you want to keep all the records from Table2


talk is cheap, supply exceeds demand
takama13
Contributor II
Contributor II
Author

Thank you, seems that is the solution.

But I now have an other issue : using this script with my full database (800,000 lines) seems to not work for some values.

If I reduce my database, more values are corrected.

I fear that the intervalmatch function creates a too large table which reach a limit (which one ?).

Would there be an alternative way to correct my values ?

Gysbert_Wassenaar

Try changing the To dates in your interval table to DayEnd(To) as To so they cover the entire day. In other words 01/04/2017 which is now 01/04/2017 00:00:00 becomes 01/04/2017 23:59:59.




talk is cheap, supply exceeds demand
takama13
Contributor II
Contributor II
Author

That does not change the result.

The values that are not changed are not necessary at the bounds.

Gysbert_Wassenaar

Try to isolate a case that is giving an incorrect result.  Then check that the data values are really really date values and not some sort of string value that merely looks like a date value.


talk is cheap, supply exceeds demand
takama13
Contributor II
Contributor II
Author

ok, there is something weird.

Here is an example of what did not work :

   

Service CodeNew Service CodeDate
COLCJS16-07-2018 22:54:00
COLCJS16-07-2018 12:00:00
COL 15-07-2018 19:30:00
COLCJS15-07-2018 05:05:00
COLCJS

14-07-2018 06:36:00

I went to my database and changed the date to the 15/07/2018 19:31.

With this changed, it worked (so I though of about a bad format, as you said).

Then I changed it back to the 15/07/2018 19:30, and it's not working again...

Gysbert_Wassenaar

What does that value turn into in Qlikview? In a listbox does it align left or right? If left then that would indicate it's interpreted as a string value by Qlikview.


talk is cheap, supply exceeds demand
takama13
Contributor II
Contributor II
Author

As a value (aligned to the right)

Gysbert_Wassenaar

Can you create a test case, i.e. a small qlikview document + source excel files, for this? Just enough data to have some records that work and one or two that don't.


talk is cheap, supply exceeds demand