Dates mismatch between FRED and imported data about fredr HOT 4 CLOSED

Thank you! very clear and useful, I'll fix my code right away!

from fredr.

Hi @ceschi ,

Thanks for bringing this to my attention. The inconsistency arose from poor handling of time series parsing in fredr_series. Since 691ca81, I've opted for simply returning the series values and dates a an xts object. Can you update fredr and try again?

devtools::install_github('sboysel/fredr')
fredr_series(series_id = 'FGCCSAQ027S')

from fredr.

I tried for the examples above and it seems to work, let me check more extensively though. Anyway, it seems that the format dates are displayed is now different when using quarterly frequency, from the YYYY-Q format to YYYY-MM-DD, am I correct?

from fredr.

Correct. That is a conscious choice to return a simple object containing observation dates and values without loss of data fidelity. If you want to coerce the returned object to a quarterly ts object:

x <- fredr_series(series_id = 'FGCCSAQ027S')
ts(data = x, start = c(1945, 3), frequency = 4)

or convert the existing object to quarterly using zoo::as.yearqtr

library(zoo)
index(x) <- as.yearqtr(index(x))

Notice how in the ts example, I assume that the date 1945-10-01 corresponds to the observation for 1945 Q3, while as.yearqtr simply divides the calendar year into 4.

from fredr.