Combining data from two tables when data is in different format

[AngelaB]

Hello fabulous Fabric community. I will do my best to describe my problem to see if what I'm trying to do is even possible. I have also included some sample data below to show what I mean.

I have two data sources that I'm trying to combine to be able to use them as a single field in reports. The two sources have come about because we changed how this data was being collected about 12 months ago (but can still be combined so don't worry about that). This means I've ended up with the data sitting across two tables -

- Table 1: the value is a summarised value (%) by ward (historical data prior to 2022)

- Table 2: the value is 'raw' or unsummarised data by response by ward. I then have a calculated measure where I summarise this as a percentage by facility (current data from 2022 onwards)

I would like to combine the two to allow me to display the summarised value across time in visualisations.

I suspect this will need a multi-step transformation but can't work it out.

Here is some sample data to illustrate what I mean:

Table 1. Historical summarised data

Ward	Timepoint	%MOB	%NUTR	%COG
Ward1	1	20%	50%	10%
Ward1	2	20%	0%	0%
Ward2	1	50%	30%	10%
Ward2	2	60%	20%	0%
Ward2	3	60%	20%	0%

Table 2. Current unsummarised data

Ward	Timepoint	MOB	NUTR	COG
Ward1	3	Yes	No	No
Ward1	3	No	No	Yes
Ward1	3	No	No	No
Ward1	3	Yes	Yes	No
Ward1	3	Yes	Yes	No
Ward1	3	Yes	No	No
Ward1	3	Yes	No	No
Ward1	3	No	Yes	No
Ward1	3	Yes	Yes	Yes
Ward2	4	Yes	No	No
Ward2	4	No	No	Yes
Ward2	4	No	No	No
Ward2	4	Yes	Yes	No
Ward2	4	Yes	Yes	No
Ward2	4	Yes	No	No
Ward2	4	Yes	No	No
Ward2	4	No	Yes	No
Ward2	4	Yes	Yes	Yes

I would like to be able to show data for Ward 1 across timepoints 1-3 in the same visualisation so need to get these combined somehow.

Thanks 🙂

[AngelaB]

Hello @Anonymous - thank you so much for this solution (and sorry for the delayed response, leave over the Festive period and all that).

I have implemented this using a new calculated table and it appears to have worked (no DAX errors etc), however the resultant merged table is patchy... some of the pre-2022 data has come through and some of the post-2022 data has come through, but there are gaps for both which is really odd. Any thoughts on what would cause a glitch like that?

Thanks again, Angela.

[Anonymous · Accepted solution]

HI @AngelaB,

Here is the calculated table formula to union two table and convert their records to same format, you can try to use it if suitable for your requirement.

Merged =
UNION (
    SELECTCOLUMNS (
        'Table1',
        'Table1'[Ward],
        'Table1'[Timepoint],
        "COG", 'Table1'[%COG],
        "MOB", 'Table1'[%MOB],
        "NUTR", 'Table1'[%NUTR]
    ),
    SUMMARIZE (
        'Table2',
        'Table2'[Ward],
        'Table2'[Timepoint],
        "COG",
            DIVIDE (
                CALCULATE ( COUNTROWS ( 'Table2' ), 'Table2'[COG] = "Yes" ),
                COUNTROWS ( 'Table2' )
            ),
        "MOB",
            DIVIDE (
                CALCULATE ( COUNTROWS ( 'Table2' ), 'Table2'[MOB] = "Yes" ),
                COUNTROWS ( 'Table2' )
            ),
        "NUTR",
            DIVIDE (
                CALCULATE ( COUNTROWS ( 'Table2' ), 'Table2'[COG] = "Yes" ),
                COUNTROWS ( 'Table2' )
            )
    )
)

Regards,

Xiaoxin Sheng