Write it out in English first. This is always a good idea, as it focusses your thought processes. For example, is the break every third row, or the tens digit of NUMBER, or something else? Once you've figured that out, and documented it, you will have your transformation (source to target mapping) specification.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
1) Take the input into a transformer.
2) Make a column tens_place in the output link.
3) The derivation of that column should be input.NUMBER/10, (Cast it as an integer column)
4) Take an aggregator and group on the basis of ID,VALUE and tens_place.
5) Take out the maximum and minimum value for NUMBER.
The break can be at any point, it should generate a new row with minimum and maximum value for every sequence stream. if any of the next value is greater than the next sequence value (like getting 8 after 5 followed by 9,10,11)) then there should be another row having the minimum as 8 and maximum as 11. Its like getting the range of all the sequence breaks for VALUE column. I have changed the example above to be understand properly.
And if there is a case where we get like 5,8,11 (single value breaks) then new row with minimum and maximum value will be same like 5|5,8|8,11|11.
then if I group it in aggregator on ID,VALUE, and SEQ then I can get min and max values for each break...If anyone can suggest me if I am going right and how can I get the above format for SEQ column that will be great.