Changelog

include the possibility to specify variable data type, which will result in a column of that type
include header into setFormat() again, which enables providing tables where the column names are in the header, where they will be spliced into the table.

introduction of the function setGroups(), which allows to summarize columns and/or rows according to a function. The positions and summarise function here are provided via the new function .sum().
now each getter ensures that it only gets things from an already validated schema.
for filter and group slots of a schema there is now a new way implemented for storing locations. Each location statement is now a list that contains either find, group or position parameters.

allow several filters to be applied in a separate function and thus for separate columns.
allow filtering on columns

allow .finding based on several character strings (that can themselves be regular expressions).

implement code so that duplicated observations in the input table are by default summarized (with a warning)

fix a bug that would not allow to register of two wide identifying variables first the wider, then the narrower wide identifying variable.

fix a bug that would not allow tables that have only one row and implicit variables to be properly reorganized.

incl. handling of flags (character additions to the numeric values of observed variables), which are now extracted into an ancillary column.

fix a bug where the header was not properly recognized when using setFilter().

introduction of getters that fulfill two functions: 1) check up on the schema one is currently building, 2) internal use to streamline the code-base
less cryptic naming of the example files that they can be understood without much explanation
complete revisit of the code-base
removal of the function setHeader(), as these information are now automatically extracted from the functions
renaming of the argument row = in setObsVar() to top =, because this is represents a clearer meaning of what this argument describes
.find() now also allows functions that evaluate to TRUE/FALSE
add a column specifier to .find().

new function setFilter to provide which columns/rows should be ignored by a schema.
handling of relative positions was changed internally to simplify the code.

generalise find_col and find_row to .find and handle it based on the context it’s called in.
include the option to specify clusters that are members of irregular groups of clusters associated to a variable that shall be recorded itself, setCluster(..., group = "grouping_variable", member = c(1, 1, 2, ...)).

include functions find_col and find_row to determine columns and rows "on the fly"

include functions setCluster, setHeader,setFormat, setIDVar and setObsVar for easier schema setup.
improved documentation
first steps in implementing errors/warnings when setting up a schema that will likely not result in a successful reorganisation

include a control that user-provided “.” as decimal symbol is turned into “[.]” to be used as regular expression.
re-enable completion of when values in the header are missing. Values are now completed with their left neighbour, as this is the most frequent use-case when values in the header are omitted.

enable the field ‘merge = character(1)’ for identifying variables, which would be used when a variable is spread over several columns that need to be combined.

initial release to CRAN

tabshiftr 0.5.0 - variable types