Errors on the Import of results of a process with a CSV file

[Xfolchf]

Describe the bug
This bug is a detail of incidents detected in a test of the process of loading data of the results progress data of the PAM process.

This test was carried out on the STAGING server, with the import file "projectmonitor_20180409172508.csv" (passed to EXCEL format), and attached to this bug report, and with the import result file "log-import-PAM .txt " extracted from the email received, also in the
annex.

These files are available under request to the authorized people.

The log file received has been analyzed starting from the 2nd record up to the end of the file on the record 3803 (the 1st record is the register of the names of headers of columns).

The anomalies found were:

1.- Detects false error in the "title_ca" field.
The reported error is: "Title ca no pot estar en blanc" (title can not be on be blank), but this field error is reported in almost every record.

2.- Report error in the "parent" field of "Parent no pot estar en blanc" (Parent can not be blank), but there are two "parent" fields and it is not clear what it refers to, considering that a yes that is informed and other no:

parent_id - DON'T CONTENTS ANY DATA
parent_external_id - WITH DATA
3.- "result_id" field without data on any record. This error is not reported in any record.

4.- Field "decidim_category_id" not reported in any record. This error is not reported within records 2nd through 28nd, but not explanation or rule of why this is happening was found.

5.- "start_date" field ( "inici"), informed in all registers. Reported error of uninformed in all except in records 14, 27, 28, 38, 40, 41 ..... (could not be found a casuistry)

6.- "end_date" field ("fi"), reported in all registers. There is reported error in the records 9, 17, 25, .... It is observed in these registers that the end date is after the actual date of execution of the process, although register 8nd is also later and no error is reported.

7.- Fields "start_date" (start) and "end_date" (end), in record 14 "start" is greater than "end" but no errors are reported.

These questions and supposed bugs should be analyzed with the help the answers to following questions, asked to someone who participated on the design of this module (Accountability)

QUESTIONS:

A) What are the verification rules applied to the process?

B) Are they correctly applied to the code?

C) It was applied a battery of tests cases that verified the implementation of these rules one by one?

D) What strategy is followed in case of detecting error? Are the changes related to the imported data applied, or, on the contrary, is the registration that is supposedly incorrect to be ruled out?

https://meta.decidim.org/processes/bug-report/f/210/proposals/13177