...
If using StringIO/BytesIO is too much, we should be able to pass file objects in general.
Comments
- The odb.connect() method is used in a very general way at the moment:
...
- when pointing to ('') it seems to be anything
...
- but wrong (pointing to a folder would not be nicer?).
When pointing to ('mars//RETRIEVE...') it shows a strong character. However, distributing the odb_api as it is would create just problems: nobody has access to mars from outside.
Instead we have the web api to download mars data....
The odb_api should not have a link to mars in itself, the data handling should be separate from the data retrieval and some web_api/odb_api combination could be instead done keeping the two players in two different packages.
Excess in verbosity... output from odb tools and odb_api is overwealming and sometimes standard output goes into standard error...
Code Block language py odb sql "select 1,2,3" >/dev/null 000 2017-10-30 13:10:04 (I) Matching row(s): 1 out of 1 000 2017-10-30 13:10:04 (I) Skips: 0 000 2017-10-30 13:10:04 (I) Execute: 0.000122 second elapsed, 0.00012 second cpuThe example is not 100% right, because when performing a select some filtering are possible, but try odb set/ odb split/ odb mdset. The output is too much and not directed to standard output.