Recovering from an Incorrect Excel Import
Hi, Phil
We have a solution containing many projects. Development is done in English (en), and we are translating to French (fr) for a client. Your tool has been absolutely invaluable in our efforts!
A couple of months ago, we imported some translations for different projects from Excel files. A few days later, we noticed that one of our projects (Project A) had translations that were completely wrong - e.g. long French strings for button captions, etc. We fixed the translations, but about two weeks ago, we added a small number of translations to Project A and noticed that the incorrect translations had returned. We reverted the changes and decided to investigate.
We loaded the tool for Project A and noticed that the incorrect translations were showing in the tool's grid, and with a green background indicating that they were imported from Excel. Although the correct translations were in the resources files (and were loaded at run time), the grid loaded incorrect translations. As a result, making changes with the tool and saving then inserted the incorrect strings into the resources files.
But far more troubling is that many of these incorrect translations are also in the global translations database. The incorrect strings are not always set as the preferred translation, but either way they should not be there.
Yesterday, we finally noticed the source of the problem: A spreadsheet with translations for Project B was imported for Project A. These errors were saved to the project database for Project A and obviously also in the global database!
We would like to:
- Clean up the global translations database; and then
- Clean up the project database.
We don't know of a quick way to fix the global database, so we assume we'll have to find each of the strings from that spreadsheet in the global database and remove the incorrect translations. Is there another way that you would recommend for this?
The bigger problem is the project database. We have taken a look at this post on the forum, but this is more about removing unused strings, whereas we would like to remove incorrect strings that are actually in use. We think it may be necessary to delete the whole project database and then rescan, which should then pick up the existing correct translations from the global database.
How should we clean up the project database? We have tried to Look for Translations in Global Database with no success (but we haven't removed all of the errors from the global database yet). Should we first delete the incorrect translations in the grid, and then use the look in the global database function? Or is it better to delete the project database to get rid of all the errors and create a new blank project database? Will the existing ml_string() function calls in the code be picked up by the tool when creating a new project database?
Your assistance is greatly appreciated.
ISI