I am new to Talend. Could any one please help for the below use case. How to perform this use case. I
id : 01
I assume the data you have shown is the format you would expect it to arrive in? If it is, then the main body of what is needed is covered in this tutorial (https://www.rilhia.com/tutorials/dynamic-column-order).
Essentially, with every row you are getting two values; the column header and value. Read these values in from your file as column header and value. Then use tMap variables (as I have done in the tutorial) to retrieve the correct value for each output column. Since the tMap variables hold their values between rows, your first row will just have your ID, the second value will have the ID and the name, the third row will have the ID, name and age. Since we know this, you can use a tAggregateRow after the tMap to group this data by ID. We know that only the last row per ID will hold all of the data. Therefore your tAggregateRow will group by ID and return the LAST value for the name and age. If configured correctly, you will end up with single rows for each of your records.
Thanks for your reply. This is not dynamic column issue. Column names are coming as rows with values with (: ) separator.
Like wise I have millions of records.
Here ID, Name and age are column names which are in row wise with their values separated by : and want to transpose the column names as headers and below the relevant values of particular column.
Age : 32
Age : 26
ID; Name; Age
Try out what I suggested, it will work. The tutorial merely shows the kind of process you need, I followed that up with a description of what you would need to do in order to solve this scenario.
This is like normalization in informatica which is not available directly in talend.
So here is turnaround
Use tmemorizerows to store last three row values as values are repeatating after 3 rows
Then in tjavaFlex get every 3rd value from tmemorize with following code in main
Declare variable in start code part as in picture
As there will be null rows for 1st and 2nd rows from source filter those in tfilterrow
I assumed we always have these key value pairs in the same order : (id,name,age). without any skip in the keys. (i.e. there wont be any id without name and age ).
With this assumption, i written in a simple logic to solve this issue.
Go through it and revert me back. I have not used any in memory components like thashoutput/ tbufferedoutput( instead i using temp file approach)
You are right . I made the code complicated. Actually I thought of using variables of tmap but not proceeded that way.
Thanks for your valuable suggestion .