Encoding error with fredr_series_observations about fredr HOT 11 CLOSED

nothing?

from fredr.

Can you post a reproducible example of the code that triggers the error? Or the full error message?

Even if it happens randomly, that will give us something to call repeatedly on our end to try and trigger what you are seeing.

It is very hard to help without any examples.

from fredr.

No encoding supplied: defaulting to UTF-8.

I'm pretty that is an encoding message, not an error, from the httr::content call within fredr::fredr_request. We can explicitly pass utf8 to the encoding parameter of httr::content.

The actual error you are experiencing would be better diagnosed, as suggested by @DavisVaughan, with a reproducible example.

from fredr.

I agree the encoding piece is a message not an error, the Error in (function (endpoint, ..., to_frame = TRUE, print_req = FALSE) : part seems an error and it prevents the correct execution of the code. I am currently running again the script (both lin-by-line and as source(script.R) but the error is not popping up, while yesterday it did. I wonder if it is something related to release dates of some particular data series.

from fredr.

Could you please provide the lines of code that you are running?

from fredr.

as soon as I can reproduce the error and find the exact line that triggers it I will post it, as of now it seems to work

from fredr.

This seems to be the piece of code behind the error

base <- fredr_series_observations(series_id='BOGMBASE', frequency='q') %>% tbl_xts() %>% /(.,1000)

Again, the error comes and goes, it's not showing up all the times.

from fredr.

@ceschi I cannot seem to reproduce that error with fredr 1.0.0. Not sure I understand the /(.,1000) part, perhaps it was a typo. It seems like you're using this line as part of a more complicated script or function. If the issue persists, try to use the debugger with debug(fredr_series_observations) and then rerun the code in question.

@DavisVaughan Perhaps related to the response timeout issues addressed in #70 ?

from fredr.

@sboysel probably, retrying like we do now might end up helping the finicky-ness

from fredr.

@sboysel Same version I am using. The /(.,1000) part is just to divide everything by 1000 from millions to billions. Last time I tried it took like 12 tries before having the error again. It is exactly part of a bigger script that collects a number of series. I'll try with the debugger then, thanks!

from fredr.

@ceschi You could also try installing the development version of fredr that as of #70 , will retry GET requests up to 10 times. This can be done with devtools::install_github('sboysel/fredr').

from fredr.