You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently tsibble only supports native date-time classes and tsibble's internally defined classes (yearweek, yearmonth, yearquarter). The problem is only packages from tidyverts recognises these date-time classes. For example, ggplot2 does not recognise yearmonth natively without specifying scale_x_yearmonth. Furthermore, as yearmonth/yearquarter stores values as integers internally rather than fractions of the year (year plus 0 for January, 1/12 for February, 2/12 for March and so on), it is very hard to convert tsibble to ts or zoo objects and makes it prohibitive to interoperate with other time series packages.
What's better to expect:
tsibble works just like tibble without additional caution (works with ggplot2 and other tidyverse operations smoothly) only with enhanced time series features.
tsibble works with zoo::yearmon or zoo::yearqtr as time indices. This will make tsibble much more versatile and interoperable.
It would be even better if tsibble provides functions to coerce to and from non-tidy time series class (e.g. ts).
The text was updated successfully, but these errors were encountered:
as_tsibble() handles the coercion of ts to a tsibble, and as.ts() vice versa. The {tsbox} package provides more functions to coerce between time series classes.
Currently
tsibble
only supports native date-time classes andtsibble
's internally defined classes (yearweek
,yearmonth
,yearquarter
). The problem is only packages from tidyverts recognises these date-time classes. For example,ggplot2
does not recogniseyearmonth
natively without specifyingscale_x_yearmonth
. Furthermore, asyearmonth
/yearquarter
stores values as integers internally rather than fractions of the year (year plus 0 for January, 1/12 for February, 2/12 for March and so on), it is very hard to converttsibble
tots
orzoo
objects and makes it prohibitive to interoperate with other time series packages.What's better to expect:
tsibble
works just liketibble
without additional caution (works withggplot2
and othertidyverse
operations smoothly) only with enhanced time series features.tsibble
works withzoo::yearmon
orzoo::yearqtr
as time indices. This will maketsibble
much more versatile and interoperable.tsibble
provides functions to coerce to and from non-tidy time series class (e.g.ts
).The text was updated successfully, but these errors were encountered: