<< | Index | >>
---
Since September the substances are no longer parsed correctly from the Packungen.xls. Probably related to the changed format in the XLS.
Running bundle exec rake spec show no failures.
Running bin/oddb2xml took less than 4 minutes.
Complaint was
Als ich das letzte Mal die Daten gezogen habe (im September), waren beim XML oddb_products.xml pro Artikel (GTIN) die Substanzen direkt vorhanden (<swissmedicsubstance>). In meinem neuen File von letzter Woche fehlen die meisten Einträge. Muss ich zusätzlich eine Option eingeben? Stellt ihr diese Daten nicht mehr zur Verfügung? Oder habe ich einfach Pech beim Runterladen gehabt? Werde ansonsten morgen alles nochmals ziehen. Habe alles mit „oddb2xml -a nonpharma“ gezogen.
Running bundle exec rake spec deleted all oddb*.xml files in my directory. Not nice!
We have 808 SubstanceSwissmedic elements out of 14737 PRD in oddb_product.xml. When just extracting the Packungen.xls I find 7149 non empty substance_swissmedic. Also I have 7149 ean entries.
Continued work on friday:
Added and option --skip-download. But time-consuming action is elsewhere. Trying to find it.
Using a version 0.9.5 of the spreadsheet gem proved to be a bad idea and broke the output.
Pushed commit Support (only) the new swissmedi package format. I did not try to support the old and the new format! Also all (or at least) most of the substance-numbers have changed. I am not sure if this was intended by swissmedic or if it is a consequence of having changed a lot of names to latin.
I will review the changes to support the no-download option next week. They don't speed up the whole process a lot because loading/parsing the big xml files is more time consuming. But no-download allows to test/improve the code offline which might be a worthwile goal in itself.