Comparing two Datasources for differences

[#1]

I have two data sources, one from Teradata and the other from Excel.   All column names are aligned across these two data sets.  I am loading both sources into my Dashboard and as expected QV is Joining them together.

If I was working in one database I could execute a Union All on these data sources and Group By every column with a Count to determine which records don't exist in both sources, i.e. count = 1, rather than 2.   Wondering if there is a simple solution to this use case utilizing Qlikview?

[Anonymous]

you can use Qualify for the data loading.

And use Unqualify for the common dimension (attributes) to compare figures

[maxgro]

I think you can do the same thing (or get the same result) in Qlik,see attachment

[mjayachandran]

In ur load script for each table, have an extra column which can identify the table from which u r loading the data,

something like

TableA:

Load

'TableA' as Tablename,

A,

B,

C

From xyz

TableB:

Load

'TableB' as Tablename,

A,

B,

C

From xyz

Now in ur dashboard have all the columns in a pivot table with count of , you can easily identify in which table what data is missing

[#5]

Hi Massimo,  Thanks for the response; that looks like what I am trying to accomplish!.  Unfortunatly I'm using the free desktop client and can't open your .qvw.  We are still evaluting purchase of QV license.  Can you please supply a screen shot of your object expression.

Many Thanks,

Lew

[maxgro]

this is the script; I did all in the script, trying to replicate your sql in database

the object is a table box, just fields, no dims or expressions

teradata:

load * inline [

a,b,c

1,2,3

4,5,6

7,8,9

a,b,c

];

excel:

NoConcatenate load * inline [

a,b,c

1,2,3

4,5,6

7,8,9

x,y,z

a,b,d

];

t:

load *, 't' as type resident teradata;

Concatenate (t) load *, 'e' as type resident excel;

DROP table teradata, excel;

g:

load

a,b,c, count(type) as cnt, concat(type) as type

Resident t

group by a,b,c;

DROP Table t;

[#7]

Hi Mahesh, My load is just as you have described, just not sure what the expression would be that would group by all columns to get the count of duplicates.