Different executing time of get_inat_obs_user

[TlaskalV]

Hello,
I found out that time needed to download per user observations differs largely between users. Do you know the reason? Is it due to different geographic storage of the data or is it restricted by API request number (e.g. issue here)?
Is there any workaround to access single user observations faster?

Below are few tests, why it takes so long for friel with only 300?

Time to download 300 observations of the user friel
> system.time({ table <- rinat::get_inat_obs_user(username = "friel", maxresults = 300) })
user system elapsed
3.92 0.07 94.66

300 observations for zdekanovkov
> system.time({ table <- rinat::get_inat_obs_user(username = "zdekanovkov", maxresults = 300) })
user system elapsed
0.7 0.0 14.8

1500 observations for zdekanovkov - still shorter than for friel
> system.time({ table <- rinat::get_inat_obs_user(username = "zdekanovkov", maxresults = 1500) })
user system elapsed
0.71 0.00 15.67

respectable 50 000 observations for finatic (no. 3 user in total, note the time which is comparable to the first result)
> system.time({ table <- rinat::get_inat_obs_user(username = "finatic", maxresults = 300) })
user system elapsed
4.17 0.08 109.93

[stragu]

Hi @TlaskalV

Thanks for the report.

Looks like the time needed is proportional to how many total observations the user has. The maxresults argument only limits how many observations will be returned by the function, but the version of the API we are using still looks for the totality of the user's observations when queried.

I think this might be fixed when we move to the more recent API, we'll have to see.

I'll keep this open, at least until we move to the newer API.