Closed JulioV closed 1 year ago
Hi @JulioV, sure thing. It looks like local_start_date_time
and local_end_date_time
are only present in Fitbit sleep summary data, so if we pass cols(local_date_time = col_character(), local_start_date_time = col_character(), local_end_date_time = col_character())
to col_types
when the device type is Fitbit, the following warning message is still produced during rule fitbit_readable_datetime
for all other Fitbit data types:
Warning message:
The following named parsers don't match the column names: local_start_date_time, local_end_date_time
To resolve all warning messages, I think we can:
cols()
to col_types
if the device type is not Fitbitcols(local_date_time = col_character())
to col_types
if the device type is Fitbit and sensor input is not sleep summary data cols(local_date_time = col_character(), local_start_date_time = col_character(), local_end_date_time = col_character())
to col_types
otherwiseI just opened a pull request (#205) to address this issue. Could you please review whenever you have some time and let me know if you'd like me to make any changes? Thank you!
Good catch, yes, thank you!
A comment in this thread reported the following warning:
This happens because those columns appear in the col specification of
read_csv_chunked
here but such columns only exist for Fitbit data and not for phone sensors.We load those columns as character because we convert them to date time with a specific time zone in this function. Thus, to avoid that warning we should pass
cols()
tocol_types
whendevice_type
is notFitbit
andcols(local_date_time = col_character(), local_start_date_time = col_character(), local_end_date_time = col_character())
when it isFitbit
.@jenniferfedor Can we implement this improvement in our next release please?