Logo Questions Linux Laravel Mysql Ubuntu Git Menu
 

Using regex in filter in Julia

It is possible to filter items that fits a simple condition to match strings in Julia:

y = ["1 123","2512","31 12","1225"]
filter(x-> ' ' in x, y)

[out]:

2-element Array{String,1}:
 "1 123"
 "31 12"

But how do I get the reverse where I want to keep the items that doesn't match the condition in a filter?

This syntax isn't right:

> y = ["1 123","2512","31 12","1225"]
> filter(x-> !' ' in x, y)
MethodError: no method matching !(::Char)
Closest candidates are:
  !(::Bool) at bool.jl:16
  !(::BitArray{N}) at bitarray.jl:1036
  !(::AbstractArray{Bool,N}) at arraymath.jl:30
  ...

 in filter(::##93#94, ::Array{String,1}) at ./array.jl:1408

Neither is such Python-like one:

> y = ["1 123","2512","31 12","1225"]
> filter(x-> ' ' not in x, y)
syntax: missing comma or ) in argument list

Additionally, I've also tried to use a regex:

> y = ["1 123","2512","31 12","1225"]
> filter(x-> match(r"[\s]", x), y)
TypeError: non-boolean (RegexMatch) used in boolean context
in filter(::##95#96, ::Array{String,1}) at ./array.jl:1408

Beyond checking whether a whitespace is in string, how can I use the match() with a regex to filter out items from a list of strings?

like image 655
alvas Avatar asked Feb 06 '23 03:02

alvas


1 Answers

In order:

  1. filter(x-> !' ' in x, y). The precedence is wrong here. The error message is telling you that it's trying to apply the ! function to a single Char argument: (!' ') in x. You need explicit parentheses:

     julia> filter(x-> !(' ' in x), y)
     2-element Array{String,1}:
      "2512"
      "1225"
    
  2. filter(x-> ' ' not in x, y). not isn't a keyword in Julia.

  3. filter(x-> match(r"[\s]", x), y). The error is telling you that it expected a boolean value but didn't get one. Unlike Python, Julia doesn't have "truthy" values. So instead of match, use contains.

    julia> filter(!contains(r"[\s]"), y)
    2-element Vector{String}:
     "2512"
     "1225"
    
like image 50
mbauman Avatar answered Feb 07 '23 17:02

mbauman