topic Re: Column Profiling Chart in App Development

Column Profiling Chart

Anonymous — Sat, 23 Sep 2017 15:23:08 GMT

Hello Everyone!

Is it possible to have a column data profiling on one chart? What I am trying to do is give a summary on the columns data showing :

1- Count of DISTINCT values

2- Count of Duplicated Values

3- Count of NULL values

For example, say I have a column "DateofBirth", I'd like to have a bar chart (or pie chart) showing the above measures. Now I have applied the above measures but when it comes to selecting a dimension (since I have to select) it messes what I am trying to accomplish

Thanks & your feedback is really appreciated.

Re: Column Profiling Chart

rubenmarin — Sat, 23 Sep 2017 16:52:50 GMT

Hi feras, and what you try to accomplish? if you want to ignore all selections you can add {1} to expressions:

Count({1} FieldName)

Re: Column Profiling Chart

Anonymous — Sun, 24 Sep 2017 11:04:35 GMT

Hi Ruben, thanks for your reply. I am not trying to ignore all selections, Say I have column Age which has the following values:

NULL

I would like to have a ONE chart that represent the following:

Duplicates Count: 3

Distinct count: 5

Null count: 2

Re: Column Profiling Chart

rubenmarin — Sun, 24 Sep 2017 11:25:31 GMT

Maybe?...

Duplicates Count:

Sum(Aggr(If(Count(Age)>1, 1),Age))

Distinct count:

Count(DISTINCT Age)

Null count:

Sum(If(IsNull(Age), 1, 0))

Re: Column Profiling Chart

Anonymous — Sun, 24 Sep 2017 11:40:37 GMT

That's fantastic; however, I am still facing the same issue, in case of a bar chart if the dimension is AGE then it will count the above measures based on that dimension. What I am trying to accomplish is that above measures in total (without going into detail)

Re: Column Profiling Chart

rubenmarin — Sun, 24 Sep 2017 15:36:04 GMT

Can you upload a sample to check?

Re: Column Profiling Chart

Anonymous — Mon, 25 Sep 2017 05:16:59 GMT

OK....

the table is:

CustomerID, AGE

1 , NULL

2 , NULL

3 , 55

4 , 55

5 , 53

6 , 52

7 , 20

8 , 20

How can I have a bar chart to display the above data as follows:

3 BARs

BAR1: 2 // NULLS

BAR2: 4// DISTINCT Values

BAR3: 2 //DUPLICATES

Hope it is clear now.

Re: Column Profiling Chart

rubenmarin — Mon, 25 Sep 2017 06:35:19 GMT

PFA, you can see the 3 bars with the expressions I posted before.

Re: Column Profiling Chart

Anonymous — Mon, 25 Sep 2017 06:37:55 GMT

I do have to select a dimension, don't I? if I Select age as the BAR chart's demsnion this will cause the chart to show the measures you wrote based on every single value... Which is not what I want.

Re: Column Profiling Chart

rubenmarin — Mon, 25 Sep 2017 06:42:44 GMT

Hi feras, this is why I'm asking for a sample or a detailed step of what you want, What means "Select Age as bar chart dimension"?

Setting all values of AGE and make the expressions for all the values? the count distinct will be always '1', what else should count?

Re: Column Profiling Chart

rubenmarin — Mon, 25 Sep 2017 06:56:50 GMT

Having all values of a dimension the count of duplicates can be:

Sum(Aggr(Count(AGE),AGE))

But I don't get how the nulls it should be counted having dimension values, or why count distinct.

Re: Column Profiling Chart

rubenmarin — Mon, 25 Sep 2017 07:45:49 GMT

Or this attachment?

There are many options if I start to imagine what you want, and english is not my native language... a sample, or at least a screenshot of what behaviour you search will be nice.