Hadd parallelization: Sources and Target have different compression settings

[fballi86]

Hi,
i have a similar problem as this one, which was never solved:
https://root-forum.cern.ch/t/hadd-parallelization-does-not-respect-compression-settings/60505/4

hadd  -j 8 "$j".${tag}.root dir*/*.root

I get:

Sources and Target have different compression settings
hadd merging will be slower

which makes my merging impossible as it’s too long.

ROOT Version: 6.34.04
x86_64-el9-gcc13-opt

Cheers,
Fabrice

[Danilo]

Hi Fabrice,

Couldn’t the setting of the target file be adapted? I apologise in advance if I do not understand the full context of the issue you are tackling.

Cheers,
D

[fballi86]

Hi @Danilo
when I do e.g.

hadd -ff -j 2 -d . $out dir*/*.root

here is what I get:

Parallelizing  with 2 processes.
hadd Target file: mc16_13TeV.604498.PhPy8EG_A14_WZqqll.e8552_s3126_r11165_p3663.root
hadd compression setting for all output: 101
Each process should handle at least 3 files for efficiency. Setting the number of processes to: 2
hadd Source file 1: dir_1_1/outTree.root
hadd Source file 2: dir_2_2/outTree.root
hadd Source file 1: dir_4_4/outTree.root
hadd Source file 3: dir_3_3/outTree.root
hadd Sources and Target have different compression settings
hadd merging will be slower
hadd Target path: ./partial0_9afa10ac-0e12-11f0-a27d-59aa9e86beef.root:/
hadd Sources and Target have different compression settings
hadd merging will be slower
hadd Target path: ./partial1_9afa10ac-0e12-11f0-a27d-59aa9e86beef.root:/
hadd Source file 1: ./partial0_9afa10ac-0e12-11f0-a27d-59aa9e86beef.root
hadd Source file 2: ./partial1_9afa10ac-0e12-11f0-a27d-59aa9e86beef.root
hadd Sources and Target have different compression settings
hadd merging will be slower
hadd Target path: mc16_13TeV.604498.PhPy8EG_A14_WZqqll.e8552_s3126_r11165_p3663.root:/

I thought using the -ff option would solve this.
Also checking the compression settings of the inputs, they seem to be 101, just like the target compression settings here.
I used a different ROOT version to create the inputs, that said. Could the issue be coming from that ?