labelr News

labelr 0.1.9

• fixed smerge() bug introduced by Base R changes to set operation behavior.

labelr 0.1.8

• fixed tabl() bug affecting use of wide.col option. Bug was introduced with the labeled.data.frame construct in 0.1.7 and affected internal call to stats::reshape().

labelr 0.1.7

• added the labeled.data.frame class, including an as_labeled_data_frame() function that assigns that class to a data.frame. Now any data.frame that is assigned frame, name, or value labels is assigned the class labeled.data.frame (while continuing to inherit the data.frame class).

• added a [ method so that subsetting and filtering of labeled.data.frame objects preserves labelr label attributes. Note that, unlike with Base R data.frames, this labeled.data.frame [ method always returns a data.frame.

• modified as_base_data_frame() so that labeled.data.frames are quietly coerced to Base R data.frames, unlike other augmented data.frames, which are (still) noisily converted to Base R data.frames.

• improved ability of add_lab_col*() and use_val_lab*() functions to detect quantity-labeled variables that are no longer numeric.

• bug fixed in greml() and gremlr() so that these handle cases where character vector passed to x argument is a single element (1L).

labelr 0.1.6

• added two convenience functions: fact2char() converts all factor data.frame variables to character variables, and irregular2() converts all irregular (see check_irregular() and irregular2v()) values in a data.frame to some other single value (NA, by default). irregular2().

• ssort() code is streamlined (and no longer issues a warning when it recycles the descending argument).

• added fact.to.char (defaults to FALSE) and irreg.to.na (defaults to FALSE) arguments to as_base_data_frame().

• removed extraneous code from transfer_labs().

• Corrected argument specification error in example flab() and slab() calls in the Introduction vignette (call was supplying argument “Female” when it should have supplied “F”) . Vignette call syntax error was introduced in v1.0.5 and was limited to the syntax of the specific example calls in the vignette. The functions themselves were not at fault and are, thus, unchanged.

labelr 0.1.5

• Added function v() to allow passing of an “unquoted character vector” (of labels or column names).

• Updated make_demo_data() and associated gender identity label examples to be more inclusive.

• Fixed some typos and copy-editing errors in documentation.

labelr 0.1.4

• Functions has_val_labs(), has_avl_labs(), has_m1_labs(), and has_quant_labs() added to facilitate logical check of a single data.frame variable (column) for the presence of, respectively: (1) any sort of value labels, (2) add_val_labs() -style value labels specifically, (3) add_m1_lab()-style value labels specifically, or (4) add_quant_labs()-style value labels specifically.

• Function all_uniquev() (alias all_univ()) added. This checks where the length of a vector is equal to the length of the unique values in the vector (is equivalent to length(x) == length(unique(x)). It is anticipated that this will be used internally in future iterations of labelr.

• Improved add_val_labs() (and add_val1()) to better detect and prevent their use on variables that already have add_m1_lab() -style many-to-one values labels. Previously, it was possible in some cases to modify select add_m1_lab()-style value labels using add_val_labs(). Now, add_val_labs() (and add_val1()) will detect and prohibit this behavior, redirecting the user to use add_m1_lab() for such variables. This reinforces the distinction between add_val_labs()-style (one-to-one) and add_m1_lab()-style (many-values-to-one-label) value labels, whereas prior behavior blurred this line.

• Modified select *1() functions (e.g., add_lab_col1(), use_val_lab1()) so that they explicitly dis-allow indirection (e.g., passing an arbitrarily named character vector containing column names) and instead require that the supplied variable name (the var argument) be the literal name of a single variable present in the supplied data.frame. The functions permit users to pass that var name unquoted or quoted, although the documentation advises the user to supply unquoted variable names.

• Two shared, vestigial code chunks removed from the following: use_val_labs(), use_val_lab1(), add_lab_cols(), add_lab_col1(), and val_labs_vec() documentation. Code involved checking for all values NA from an earlier iteration of use_val_labs(). Code is more concise; functions, arguments, and outputs are unchanged.

• val_labs_vec() now strips any lingering attributes from the returned character vector.

• Improved handling of errant and varying var arg specifications in val_labs_vec(), add_lab_col1(), and use_val_lab1() to be more flexible and informative if a non-existent var is supplied by the user.

• Move internal functions to the beginning of the the body of add_val_labs(), add_val1(), add_m1_lab(), and add1m1().

• Copy edit fixes and improved argument explanations in documentation items (e.g., an outdated comment in README, an infelicitous example in use_val1() and add_lab_col1() documentation).

• function get_val_lab1() (with alias gvl1()) added, providing a variant of get_val_labs() that follows the conventions of other *1() functions (i.e., allowing only a single variable, whose name may be passed unquoted to the var argument). In addition, get_val_lab1() includes a simplify argument (FALSE by default) that allows the user to return the value-label mapping as a named vector.

labelr 0.1.3

• functions add_lab_dummies() and add_lab_dumm1() (with aliases ald() and ald1()) added. These generate a dummy variable for each label of a value-labeled variable and return the supplied data.frame with these dummy variables added.

• functions factor_to_lab_int() (with alias f2int()) and lab_int_to_factor() (with alias int2f()) added to allow easy conversion of labeled integer variables to factors and of factors to labeled integer variables.

• function axis_lab() (with alias alb()) added - helper that allows one to easily use variable name labels as y or x axis labels with Base R plots or ggplot2.

• function all_quant_labs() (with alias allq()) added: automatically add quantile-based numerical range value labels for all numeric variables that meet specifications.

• function tabl() updated. labs.on = TRUE is now the default. qtiles argument added that allows for non-value-labeled, many-valued numeric variables to be converted temporarily and on the fly to quantile category variables, so that they can be included in returned table results. Tabulation “engine” now relies on base table(), which makes select tabulations much faster.

• function use_val_lab1() (with alias uvl1()) added. This provides the same functionality as use_val_labs(), but allows the user to supply only one unquoted variable, whereas use_val_labs() requires quoted variable name arguments but allows the user to pass a vector of multiple variable names in one call.

• function add_lab_col1() (with alias alc1()) added. This provides the same functionality as add_lab_cols(), but allows the user to supply only one unquoted variable, whereas add_lab_cols() requires quoted variable name arguments but allows the user to pass a vector of multiple variable names in one call.

• updated add_name_labs(), drop_name_labs(), convert_labs(), and clean_data_atts() functions, so that changes (add or drop) of name.labs attributes are applied to native labels() attribute and, in case of convert_labs(), so that any incoming variable/column level labels() attributes (e.g., column name labels from a haven-imported tibble) are converted to labelr name.labs. Primary benefit of this is that RStudio View() will now show name labels underneath column names.

• use_val_labs(), add_lab_cols(), and val_labs_vec() fixed to convert any irregular character values to NA labels (was converting to “NA”).

• tabl() wide.cols (pivot, cross-tab) functionality improved to show 0 instead of NA for empty cross-tab cell counts.

• add_quant_labs() and add_quant1() now round non-integer auto-generated quantile value labels to avoid long, repeating-decimal value labels (e.g., “q067” instead of “q066.66666…”).

• updated ssort() to preserve rownames.

• updated gremlr(), greml(), slab(), sbrac(), somel(), taill(), sfilter(), and ssort() to handle situations involving (or producing) one-column data.frames.

• Defensive programming improvements made to various value-label-related functions, so that they respond more gracefully and informatively to non-valid variable name arguments.

• split up original one vignette into two respective vignettes: Introduction and Special Topics.

• light copy-editing updates to README, vignette, and documentation.