import synoptic
df = synoptic.TimeSeries(...).df
df.synoptic.pivot()
The default DataFrame is in long format; one row for each unique observation
Requesting a obtimezone='local'
still returns column of dataframe in UTC time. This is because Polars datetime type is timezone aware, so you can't have multiple timezones in the same column.
If you need all data in their local timezone, then use df.partition_by('stid')
- Station elevation has units of feet.
- Sensor position has units in meters.
I don't allow the use of the legacy API behavior by not letting the user omit the pmode
argument. Its default is "totals"
The QC column is always provided, but does not mean QC checks were performed.
-
Data is provided in long format; one unique observation per row.
-
Only float values are parse in the
value
column. Variables that return string values are in thevalue_string
column. Some values are complex structs, and I haven't implemented those yet. -
All column names are changed to lower case.
-
Don't do anything with
"SENSOR_VARIABLES"
. This tells us what variables are used to derive other variables, but I don't see how it is useful when doing any analysis.
- The presence of the
qc_flags
column means that some QC checks were applied.