Skip to content
Snippets Groups Projects

Multi input

Merged Bert Palm requested to merge multi-input into develop
All threads resolved!
  • command line now accespts multiple options -d (data input)
  • SaQC can eat now list of Data-like
  • SaQC can eat now list of Flags-like
  • duplicate columns in input will overwrite former columns (with same name) and we throw a warning see fist thread why this is like this)

i suggest no squeeze and delete for this MR, because its quite clean. also in develop only one commit will be present..

closes #257 (closed)

Edited by Bert Palm

Merge request reports

Loading
Loading

Activity

Filter activity
  • Approvals
  • Assignees & reviewers
  • Comments (from bots)
  • Comments (from users)
  • Commits & branches
  • Edits
  • Labels
  • Lock status
  • Mentions
  • Merge request status
  • Tracking
  • assigned to @palmb

  • Bert Palm changed the description

    changed the description

  • Bert Palm changed milestone to %2.0

    changed milestone to %2.0

  • Bert Palm requested review from @schaefed

    requested review from @schaefed

  • Author Owner

    We throw a warning on duplicate columns and overwerite the formaer data.

    we do this because renaming isnnt trivial.. We can implement a rename (append a number) workflow later if we have time for that..

    why this is complex

    imagine our appendum is _N, where N is the number to append.. now imagina someone provide some dataset with exaxtly our scheme, letz say, with the columns a, a_1, a_2, a_2 (yes ther is a double a_2) and a second dataset with a, b, c..

    the steps are the following

    • read all datasets columns
    • check that a_2 is duplicate name and a is a duplicate name
    • check that we cannot just append _1 to a (because this already exist)
    • somhowe straight forward we end up with a, a_1, a_2, a_2_1, a_3, b, c

    This issnt nice because we have a, b, c and a_1, a_2, a_3 and a_2_1.. which all does not belong to each other..

    why this is real

    even if one might say this won't happen, because its only is valid for people that exactly use our rename scheme, i want to point out that it becomes a thing, if people use saqc twice..

    after a first run with columns a, a all works and the result is a, a_1.. The next day with the original data (because all worked so fine) and the new data one start saqc with a, a and a, a_1 and it become ugly... (see above)

    TLDR

    no time for making it right

    Edited by Bert Palm
  • Bert Palm mentioned in issue #257 (closed)

    mentioned in issue #257 (closed)

  • Bert Palm changed the description

    changed the description

  • David Schäfer
  • David Schäfer
  • David Schäfer
  • David Schäfer approved this merge request

    approved this merge request

  • Bert Palm added 9 commits

    added 9 commits

    • 4226c601...c341115d - 6 commits from branch develop
    • 70ec5a87 - [FIX] raise informative warning if DictOfSeries is called with non-unique DataFrmae
    • 7826433a - make __main__.py take multiple data sources
    • c9f1ca4d - make SaQC take a list of multiple data/flags objects

    Compare with previous version

  • Bert Palm resolved all threads

    resolved all threads

  • merged

  • Bert Palm mentioned in commit 28eaf64d

    mentioned in commit 28eaf64d

  • Bert Palm mentioned in issue #264

    mentioned in issue #264

  • Bert Palm mentioned in issue #128 (closed)

    mentioned in issue #128 (closed)

  • Please register or sign in to reply
    Loading