The power of bash

Very often I remember the name of the R function I need to use for a particular problem but I don’t quite remember how to use it exactly… And, even worse, what I remember is that last time I had to use this function it took me a long time to figure out how to use it… Since I’m not ready to go through that learning process again, I decide to look for the R script on my computer that did make use of this function. And here comes the power of bash! (if you have bash on your computer, of course…). Well, to be more precise, I actually refering to the particular find and grep commands here that can be used in bash of course (if you’re on linux or macos), but also in any other UNIX-type shell. Let’s demonstrate its power, using my specific example.

The R function I wanted to use was sf::st_simplify(). I knew I used it recently (in the last 2 months) in an Rmarkdown file, but I couldn’t remember which one. Looking for all the Rmarkdown files in the current file hierarchy is done by

find ~/Dropbox/aaa -name "*.Rmd" -mtime -10w -type f

The find command is followed by the path of the location where we want to start the file search (here ~/Dropbox/aaa). Then it is followed by a number of options specified by pairs of options names (preceded by the - sign) followed by the option value. For example, here -name "*.Rmd" specifies that we are looking for files with names ending with the .Rmd extension, -mtime -10w means that we are looking at files that were modified less than 10 weeks ago and -type f means that we are looking at regular files (excluding directories, links, etc…). If we had been interested in file that were modified more than 3 months ago instead of less than 10 weeks ago, that would have been specified by -mtime +3m. Among these options, one is particularly amazing in the sense that it allows to run another command on the files that were selected. This option is -exec and is followed by the command we want to run. For example, here, I want to search for the st_simplify pattern in the selected files:

find ~/Dropbox/aaa -name "*.Rmd" -mtime -10w -type f -exec grep -n "st_simplify" {} \;

Here it says that, on each of the files that were selected by the first part of the find command (this files are here specified by the {} placeholder), we want to execute the grep command (with the -n option in order to also print the line number where the pattern was found in the file). Don’t forget here the \; sign that closes the value of the -exec option. Indeed, contrary to the other options of the find command that have, as a value, only one “word” (i.e. character string without any space), the -exec may have (and typically have) many words as its value. Thus, the value of the -exec command is all the words that are between exec and \;. Adding the -print option (no value) will have the effect to print the name of the file where the pattern was found:

find ~/Dropbox/aaa -name "*.Rmd" -mtime -10w -type f -exec grep -n "st_simplify" {} \; -print

And here it is!