How to clean and transform 3 columns of a csv file

[Anonymous]

Hello !

I'm a beginner on Talend but I'm so stuck that this is the first time I'm posting on a forum !
Here is the deal, I want this csv :

Host Extract_date Size Used Avail Use% Mounted_on
host1 4112017 1008M 299M 659M 32% /MIDDLE/blabla
host1 4112017 6.0G 188M 5.5G 4% /MIDDLEDATA/blabla
host2 4112017 12G 71M 12G 1% /MIDDLELOGS/blabla

to become this :

Host ExtractDate Size Used Avail UsePercent MountedOn
host1 4112017 0.98 0.29 0.64 32 /MIDDLE/blabla
host1 4112017 6.0 0.18 5.5 4 /MIDDLEDATA/blabla
host2 4112017 12 0.0612 1 /MIDDLELOGS/blabla

1) Change headers :

I can manage this in metadatas

2) Apply these rules :

- If "Size","Used or "Avail" contains the letter "M" at the end :

  a) Get rid of the M (I'm using Treplace regexp for this)

  b) Cast from String to Float (I'm using Tmap for this)

  c) Divide by 1024 (to get a size in G) (Tmap too)

  d) Round the result, 2 digits after comma (Tmap again)

- If "Size","Used or "Avail" contains the letter "K" :

  Appl the same except step c) needs to be done twice

- If "Size","Used or "Avail" contains the letter "G" :

  apply step a) and b) only

How can I merge all of these outputs ? I will have duplicate lines like this !

This works but ...

I need some help ! :'(

Greg.

[Anonymous]

Do you want to fuse all these processening steps together?