knitr::opts_chunk$set( collapse = TRUE, comment = "#>" )
nc
is a package for named capture regular expressions (regex), which
are useful for parsing/converting text data to tabular data (one row
per match, one column per capture group). In the terminology of regex,
we attempt to match a regex/pattern to a subject, which is a string of
text data. The regex/pattern is typically defined using a single
string (in other frameworks/packages/languages), but in nc
we use a
special syntax: one or more R arguments are concatenated to define a
regex/pattern, and named arguments are used as capture groups. For
more info about regex in general see
regular-expressions.info
and/or the Friedl book. For more
info about the special nc
syntax, see help("nc",package="nc")
.
Below is an index of topics which are explained in the different vignettes, along with an overview of functionality using simple examples.
Capture first is for the situation when your input is a character vector (each element is a different subject to parse), you want find the first match of a regex to each subject, and your desired output is a data table (one row per subject, one column per capture group in the regex).
subject.vec <- c( "chr10:213054000-213,055,000", "chrM:111000", "chr1:110-111 chr2:220-222") nc::capture_first_vec( subject.vec, chrom="chr.*?", ":", chromStart="[0-9,]+", as.integer)
A variant is doing the same thing, but with input subjects coming from a data table/frame with character columns.
library(data.table) subject.dt <- data.table( JobID = c("13937810_25", "14022192_1"), Elapsed = c("07:04:42", "07:04:49")) int.pat <- list("[0-9]+", as.integer) nc::capture_first_df( subject.dt, JobID=list(job=int.pat, "_", task=int.pat), Elapsed=list(hours=int.pat, ":", minutes=int.pat, ":", seconds=int.pat))
Capture all is for the situation when your input is a single character string or text file subject, you want to find all matches of a regex to that subject, and your desired output is a data table (one row per match, one column per capture group in the regex).
nc::capture_all_str( subject.vec, chrom="chr.*?", ":", chromStart="[0-9,]+", as.integer)
Capture melt is for the situation when your input is a data table/frame that has regularly named columns, and your desired output is a data table with those columns reshaped into a taller/longer form. In that case you can use a regex to identify the columns to reshape.
(one.iris <- data.frame(iris[1,])) nc::capture_melt_single (one.iris, part =".*", "[.]", dim =".*") nc::capture_melt_multiple(one.iris, column=".*", "[.]", dim =".*") nc::capture_melt_multiple(one.iris, part =".*", "[.]", column=".*")
Capture glob is for the situation when you have several data files on disk, with regular names that you can match with a glob/regex. In the example below we first write one CSV file for each iris Species,
dir.create(iris.dir <- tempfile()) icsv <- function(sp)file.path(iris.dir, paste0(sp, ".csv")) data.table(iris)[, fwrite(.SD, icsv(Species)), by=Species] dir(iris.dir)
We then use a glob and a regex to read those files in the code below:
nc::capture_first_glob(file.path(iris.dir,"*.csv"), Species="[^/]+", "[.]csv")
Helpers describes various functions that simplify
the definition of complex regex patterns. For example nc::field
helps avoid repetition below,
subject.vec <- c("sex_child1", "age_child1", "sex_child2") pattern <- list( variable="age|sex", "_", nc::field("child", "", "[12]", as.integer)) nc::capture_first_vec(subject.vec, pattern)
It also explains how to define common sub-patterns which are used in several different alternatives.
subject.vec <- c("mar 17, 1983", "26 sep 2017", "17 mar 1984") pattern <- nc::alternatives_with_shared_groups( month="[a-z]{3}", day="[0-9]{2}", year="[0-9]{4}", list(month, " ", day, ", ", year), list(day, " ", month, " ", year)) nc::capture_first_vec(subject.vec, pattern)
Any scripts or data that you put into this service are public.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.