I like the %like%
operator in data.table quite a lot as I'm used to SQL queries. In PostgreSQL there is also the ILIKE option where the __i__ stands for case-__insensitive__.
What do you think of including %ilike%
into data.table?
To include this into data.table I would create an operator %ilike%
in ilike.R based on like.R adding ignore.case = TRUE
to the grep()
calls as follows:
ilike <- function(vector, pattern)
{
# Intended for use with a data.table 'where'
# Don't use * or % like SQL's like. Uses regexpr syntax - more powerful.
if (is.factor(vector)) {
as.integer(vector) %in% grep(pattern,levels(vector), ignore.case = TRUE)
} else {
# most usually character, but integer and numerics will be silently coerced by grepl
grepl(pattern,vector, ignore.case = TRUE)
}
# returns 'logical' so can be combined with other where clauses.
}
"%ilike%" = ilike
reproducible example:
require(data.table)
cars = data.table(cars = rownames(mtcars), mtcars)
cars[ cars %like% 'fiat' ] # no case-insensitive search possible
cars[ grep('fiat', cars, ignore.case = TRUE) ] # using comparably long grep
cars[ cars %ilike% 'fiat' ] # the new %ilike%
You could use the (?i)
tag instead:
cars[cars %like% '(?i)fiat']
True that works as well, wasn't aware of such a regex. Nevertheless %ilike%
seems a more intuitive to me. Maybe because I know PostgreSQL. Overall this is a question of coding style anyway.
@andreasLD thanks again for raising. You probably saw since you commented there as well but noting for the record that this was closed in #3333 and #3552
Most helpful comment
True that works as well, wasn't aware of such a regex. Nevertheless
%ilike%
seems a more intuitive to me. Maybe because I know PostgreSQL. Overall this is a question of coding style anyway.