How can I replace many words in a string, using Haskell?

Question

Let's say I have a string I want to modify:

str = "I have several paints. For example, red, green, and blue."

I want to turn it into: I have several paints. For example, XXX, XXX, and blue.

So I have a list of words to replace with XXX:

colors = ["red", "green"]

I want to replace the colors from colors in str. Here's what I have so far:

First, a function to replace something with "xxx": subColor pat str = subRegex (mkRegex pat) str "xxx"
Next, I'm trying to map that or fold it over the list of colors. But I'm stuck: map just returns copies of the string, each with a single word replaced, and I've been getting errors with foldl. What am I doing wrong?

Edit: I should also say that I will sometimes have multi-line words, so using words (as suggested in another similar question's answers) isn't really practical for me.

Edit2: I should add that one big regex pattern, like blue|red probably wouldn't work for me either, since I actually have several thousand of these things to replace.

Why does "several thousand things to replace" imply "one big regex pattern wouldn't work"? — Daniel Wagner
– Daniel Wagner, Commented Aug 22, 2019 at 4:29
Really really looks like a foldl(/foldr) type of problem. With your order of the arguments in subCOlot (pat before str) I guess foldr just fits? foldr subColor str colors — dvaergiller
– dvaergiller, Commented Aug 22, 2019 at 12:54

luochen1990 · Accepted Answer · 2019-08-22 06:28:54Z

3

What you want isn't map over several pats and get a list of new strs, but apply subColor pat for each pat and combine these functions and apply it to str:

subColors :: [String] -> String -> String
subColors pats str = foldMap (Endo . subColor) pats str

In above code, the foldMap use the instance Monoid (Endo a), which is same as foldr (.) id.

And, I still suggest you try the "one big regex pattern" solution, since a regex engine should do more optmization than you do, but that depends on the implementation of the regex engine, so if performance matters, you need to have a test first.

answered Aug 22, 2019 at 6:28

luochen1990

3,9051 gold badge25 silver badges42 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Joseph Sible-Reinstate Monica Over a year ago

Why complicate it? Why not just foldr subColor pats str?

luochen1990 Over a year ago

@JosephSible You mean foldr subColor str pats ? That is a good idea !

Joseph Sible-Reinstate Monica Over a year ago

Yep, I got them backwards. That's what I meant.

Collectives™ on Stack Overflow

How can I replace many words in a string, using Haskell?

1 Answer 1

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related