tsibble v0.4.0

Breaking changes

• Restructured windowed functions to follow the purrr style exactly (#35):
  • slide(), tile(), stretch() return lists only instead of numerics before.
  • added slide2(), pslide() to map over multiple inputs simultaneously.
  • added other type-stable variants such as slide_dbl(), slide_int(), slide_chr(), slide_lgl().
  • A negative window size results in a backward moving direction.
  • slide() gained a new argument .partial to support partial sliding.
• Renamed x to .x in slider(), tiler(), stretcher().
• pslider(), ptiler(), pstretcher() support multiple inputs now, and split them in parallel.

New features

• New holiday_aus() for Australian national and state-based public holiday.
• Defined diff() for year-week, year-month, and year-quarter.
• yearweek(), yearmonth(), yearquarter() supported for character.
• Added slide2(), pslide(), tile2(), ptile(), stretch2(), pstretch() to slide over multiple inputs simultaneously (#33).
• New S3 generics units_since() for index classes.
• New is_53weeks() for determine if the year has 53 ISO weeks.
• New S3 generics key_sum() for extending tsibble.

Improvements

• as_tsibble.ts() removed the tsp attribute from the value column.
• Dropped the "lst_col" attribute from lst_ts (#25).
• More informative error message when a data frame is passed to tsibble().
• More informative error message for fill_na() and count_gaps when a tsibble of unknown interval.
• as_tsibble.tbl_ts() & as_tsibble.grouped_ts() now return self (#34).
• id() is used in the tsibble context (e.g. as_tsibble(), tsibble(), build_tsibble()) regardless of the conflicts with dplyr or plyr, to avoid frustrating message (#36).
• select.tbl_ts() now preserved index.

Bug fixes

• Fixed bug in as.ts.tbl_ts() for ignoring the value argument when the key is empty.
• Fixed bug in [.tbl_ts() when subsetting columns by characters (#30).
• Fixed bug in fill_na.tbl_ts() dropping custom index class (#32).
• Fixed bug in format.yearweek() due to the boundary issue (#27).
• If a column passed as index contains NA, abort.

Misc

• Suggested on nycflights13 >= 1.0.0.

tsibble v0.3.0

The tsibble package has a hexagon logo now! Thanks Mitch (@mitchelloharawild).

New functions

• New difference() computes lagged differences of a numeric vector. It returns a vector of the same length as the input with NA padded. It works with mutate().
• Added the support of gather()/spread(), nest()/unnest() to tbl_ts.

Defunct

• retired tsummarise().

Internal changes

• If index2 deviates from index (using index_by()), the index2 will be part of grouping variables.

tsibble v0.2.0

This release (hopefully) marks the stability of a tsibble data object (tbl_ts). A tbl_ts contains the following components:

• key: single or multiple columns uniquely identify observational units over time. A key consisting of nested and crossed variables reflects the structure underlying the data. The programme itself takes care of the updates in the "key" when manipulating the data. The "key" differs from the grouping variables with respect to variables manipulated by users.
• index: a variable represent time. This together the "key" uniquely identifies each observation in the data table.
• index2: why do we need the second index? It means re-indexing to a variable, not the second index. It is identical to the index most time, but start deviating when using index_by(). index_by() works similarly to group_by(), but groups the index only. The dplyr verbs, like filter(), mutate(), operates on each time group of the data defined by index_by(). You may wonder why introducing a new function rather than using group_by() that users are most familiar with. It's because time is indispensable to a tsibble, index_by() provides a trace to understanding how the index changes. For this purpose, group_by() is just too general. For example, index_by() + summarise() aggregates data to less granular time period, leading to the update in index, which is nicely and intuitively handled now.
• interval: an interval class to save a list of time intervals. It computes the greatest common factor from the time difference of the index column, which should give a sensible interval for almost all the cases, compared to minimal time distance. It also depends on the time representation. For example, if the data is monthly, the index is suggested to use a yearmonth() format instead of Date, as Date only gives the number of days not the number of months.
• regular: since a tsibble factors in the implicit missing cases, whether the data is regular or not cannot be determined. This relies on the user's specification.
• ordered: time-wise and rolling window functions assume data of temporal ordering. A tsibble will be sorted by its time index. If a key is explicitly declared, the key will be sorted first and followed by arranging time in ascending order. If it's not in time order, it broadcasts a warning.

Breaking changes

• Deprecated tsummarise() and its scoped variants. It can be replaced by the combo index_by() + summarise() (#20). tsummarise() provides an unintuitive interface where the first argument keeps the same size of the index, but the remaining arguments reduces rows to a single one. Analogously, it does group_by() and then summarise(). The proposed index_by() solves the issue of index update.
• Renamed inform_duplicates() (defunct) to find_duplicates() to better reflect its functionality.
• key_vars() and group_vars() return a vector of characters instead of a list.
• distinct.tbl_ts() now returns a tibble instead of an error.
• No longer reexported dplyr::do() and tidyr::fill(), as they respect the input structure.
• Defunct index_sum(), and replaced by index_valid() to extend index type support.

New functions

• index_by() groups time index, as the counterpart of group_by() in temporal context.
• A new S3 generic count_gaps() and gaps() counts time gaps (implicit missing observations in time).
• yearweek() creates and coerces to a year-week object. (#17)

API changes

• fill_na.tbl_ts() gained a new argument of .full = FALSE. .full = FALSE (the default) inserts NA for each key within its time period, TRUE for the entire time span. This affects the results of fill_na.tbl_ts() as it only took TRUE into account previously. (#15)
• Renamed the drop argument to .drop in column-wise dplyr verbs.
• Dropped the duplicated argument in pull_interval().
• group_by.tbl_ts() behaves exactly the same as group_by.tbl_df now. Grouping variables are temporary for data manipulation. Nested or crossed variables are not the type that group_by() thinks.

Improvements

• Added overall time span to the glimpse.tbl_ts().
• Slightly improved the speed of fill_na().

Bug fixes

• Fixed transmute.tbl_ts() for a univariate time series due to unregistered tidyselect helpers. (#9).
• Fixed bug in select.tbl_ts() and rename.tbl_ts() for not preserving grouped variables (#12).
• Fixed bug in select.tbl_ts() and rename.tbl_ts() for renaming grouped variables.

Internal changes

• A tbl_ts gains a new attribute index2, which is a candidate of new index (symbol) used by index_by().
• The time interval is obtained through the greatest common factor of positive time differences. This covers broader cases than the minimal value.
• attr(grouped_ts, "vars") stores characters instead of names, same as attr(grouped_df, "vars").

tsibble v0.1.5

This release introduces major changes into the underlying tbl_ts object:

• Dropped the attribute "key_indices" from a tbl_ts class to reduce the object size, and computed on the fly when printing.
• Gained a new attribute "ordered" to identify if it is arranged by key and index in ascending order. If not, broadcast a warning. The warning likely occurs to arrange() and slice() functions.
• The "index" attribute in a tbl_ts object is a symbol now instead of a quosure.
• The "key" attribute in a tbl_ts object is an unnamed list of symbols.

New functions

• "key" helpers:
  • key_update() to change/update the keys for a given tsibble.
  • unkey() as an S3 method for a tsibble of key size < 2.
  • key_indices() as an S3 method to extract key indices.
• split_by() to split a tsibble into a list of data by unquoted variables.
• build_tsibble() allows users to gain more control over a tsibble construction.
• Added as_tsibble.msts() for multiple seasonality time series defined in the forecast package.

Bug fixes

• Fixed as_tsibble.ts() for daily time series (when frequency = 7).
• group_by.tbl_ts() does not accept named expressions.

Internal changes

• No longer throw an error when grouping the index.
• An interval of regularly spaced tsibble is (re)computed when creating the tsibble and performing the row-wise operations (like filter() and slice()). This avoids unnecessary re-computation for many function calls.

tsibble 0.1.3

New functions

• Added the scoped variants for tsummarise() including tsummarise_all(), tsummarise_if(), tsummarise_at().

API changes

• The windowed functions, including slide(), tile(), stretch(), are no longer defined as S3 methods. Several new variants have been introduced for the purpose of type stability, like slide_lst() (a list), slide_dfr() (a row-binding data frame), slide_dfc() (a column-binding data frame).
• The index variable must sit in the first name-value pair in tsummarise() instead of any position in the call.
• transmute.tbl_ts() keeps the newly created variables along with index and keys, instead of throwing an error before.
• Depends on purrr (>= 0.2.3)

Bug fixes

• Fixed the error message in glimpse.tbl_ts()
• Fixed format.key() for nesting crossed with another nesting.

tsibble 0.1.2

• New functions: inform_duplicates()
• Reexported functions: dplyr::*_join(), dplyr::transmute(), dplyr::distinct(), tidyr::fill()
• Bug fixes. Thanks @mitchelloharawild for reporting several bugs.
• More informative error message.
• Better coding logic for grouped_ts.

Initial release on CRAN

Initial release on CRAN

Pre-release on Github

This release is primarily for John M. Chambers Statistical Software Award 2018.