multiply concatenate

[#1]

hello,

does anyone knows if by doing multiplies concatenate like this

concatenate (t1) load ... from qvd

concatenate (t1) load ... resident x

concatenate (t1) load ... resident y

where the data sorce is from few sources and part of it is grouped by and some are not.

does the concatenate is doing distinct on the data.

for example in table y i had field named customer and when i do the concatenate (t1) load with field customer

i get all the data in t1 table .

but when i concatenate y to t1 without filed customer i get only part of the data of y in t1.

and its looks like it distinct the rows in y when its concatenate to t1.

t1 is a wide table with lot of filelds and is linked by few fields to anther table.

any help will be greate

[:D]

[#2]

Hi,

u can send me some sample data

Regards,

Sampath Kumar . G

[erichshiino]

did you do some distinct during this concatenate load?

some concatenate (t1) load distinct?

Try to include flags in the table during the load to be sure you got all the lines:

t1:

load 'original' as flag,...

concatenate(t1)

load 'from x' as flag...

resident x

concatenate(t1)

load 'from y'as flag ...

resident y...

Rgds,

[Miguel_Angel_Baeyens]

Hello,

As Erich mentions above, if you are not doing any LOAD DISTINCT you are not reducing data nor joining, that's the main difference between JOIN and CONCATENATE. CONCATENATE will append records from current table to the previous one (or the table specified usgin parentheses).

If the table T1 has fields A and B, and you concatenate T2 with fields C and D, the result will be only one table (T1) with nulls as values for fields C and D in case of records from original T1 and nulls for A and B in case of records coming from T2. You will be creating more records and adding columns to an existing table, which will have some impact in performance.

What I used to is to rename fields when concatenating (fact tables coming from different sources, for example), and create a new field so I can get values from one source or the other with set analysis or conditionals, as in the sample code below.

Invoices:LOAD I_Code AS Code, I_Amount AS Amount, I_Date AS Date, 'Invoices' AS SourceFROM Source.Invoices; Shipments: // this label will be ignored, the records from source2 below will be appended to "Invoices"CONCATENATE LOAD S_Code AS Code, S_Amount AS Amount, S_Date AS Date, 'Shipments' AS SourceFROM Source.Shipments;

I can now create a master calendar which will affect to both invoices and shipments, or have a "master" Customer dimension table, for example.

Hope that helps.

[#5]

hi guys,

thank for all the answers!!

i am using qlik for the last 7 years , so lets past the beginner problems.

i asked it just to know if some one ever saw this kind of problems.

for may exprience there is a problems with concatenate which i found in the past.

for example if you make table t which has linked filelds to anther table

and in the next concatenate you dont add all the linked fields your data will be lost.

or maybe it will be lost if you binary the file to another one.

last example is when you have syntetic keys more than 1 and they are involved with each other like you

have syn1-4 which 4 is built from fields from 1 and 2 or 3 syntetic key than if you binary the file to

you may loose data.

i know all these sound wird but belive me i so such things in my own eyes :(.

i just want to know if anyone ever has these kind of problems?

thanks

dodo