Different results based on allowed file formats

[siccovansas]

When I search rijksoverheid.nl (the main website of the Dutch government) I retrieve 384 files (60 xlsx, 323 xls, 1 unknown/xlsm) with the default set of allowed file formats (as specified in constants.py). When I add 'ods' to the list of allowed file formats and retry rijksoverheid.nl then I only retrieve 298 files (24 xlsx, 149 xls, 124 ods, 1 unknown/xlsm). I have tested both versions 3 times and the results were always like this.

So, how can it be that I retrieve less files when I add a file format?

[waldoj]

You might try running an equivalent search on Bing, searching for a series of filetypes—once with ODS, once without—and see if you can replicate the results. (This service is powered by Bing's API.) My suspicion is that this problem is on Bing's end.

[waldoj]

I just ran that search on Bing, and I get 635 results with ODS, and 521 results without. (Search engines being what they are, it would not surprise me to find out that we're being shown different results for the same queries.) So that would seem to point to a problem with LMGTDFY, and not a problem with Bing.

[siccovansas]

I get almost the same results when using that URL. With ODS gives me 635, without 519.