Hi everyone !
I should also add that I found some discrepancies in the v2_0_2 version of CExD and CED impact methods. For example there are six elementary flows named "Gas, natural, 46.8 MJ per kg, in ground" (three classified in "resource > in ground" and three other in "resource > unspecified"), with the same IDs (resp. 26132 and 26142), but with differet characterization factors.
Therefore I would currently use the "ecoinvent v3-4 lcia methods" pack until the next update !