Cleaning text to strip [] and "" in julia dataframe - julia

Question

I have a dataframe with a column text that a is a list of strings, like this:
text
["text1","text2"]
["text3","text4"]
How can I clean de string to have another column text_clean like this:
text
text1,text2
text3,text4
When I type in repl df I get:
text
String
["string"]
["string","anotherestring"]
but when I type:
df[!,:text]
I get:
"[\"string\"]"
"[\"string\",\anotherestring\"]"
I would like to create a new dolumn, called text_clean:
string
string, anotherstring
Thanks

Answer 1

julia> a = [["text", "text2"], ["text"], ["text", "text2", "text", "text2"]]
3-element Vector{Vector{String}}:
["text", "text2"]
["text"]
["text", "text2", "text", "text2"]
julia> join.(a, ",")
3-element Vector{String}:
"text,text2"
"text"
"text,text2,text,text2"
replace a with your column, like df.text

Answer 2

It seems that your strings literally have values containing [s, "s, etc.
First of all, make sure that this is intended. For eg., you might have something like a vec = ["string", "anotherstring"]. At some point before this, you might have code doing the equivalent of df[1, :text] = string(vec). Instead, do df[1, :text] = join(vec, ", ") when assigning to the text column, to have that original column itself be clean.
If the above doesn't apply, and you have to deal with the column as given, then you create your new cleaned column like this:
julia> df = DataFrame(:text => [string(["hello", "world"]), string(["this","is","SPARTA"])])
2×1 DataFrame
Row │ text
│ String
─────┼──────────────────────────
1 │ ["hello", "world"]
2 │ ["this", "is", "SPARTA"]
julia> df[!, :text_clean] = map(df.text) do str
str |>
s -> strip(s, ('[', ']')) |> #remove [ ]
s -> strip.(split(s, ", "), '"') |> # remove inner "
sv -> join(sv, ", ")
end
2-element Vector{String}:
"hello, world"
"this, is, SPARTA"
(You might have to adjust the second argument to split above based on whether or not you have a space after the commas in the text column.)
Or, making use of Julia's own syntax parsing,
julia> df[!, :text_clean] = map(df.text) do str
str |> Meta.parse |>
ex -> ex.head == :vect && eval(ex) |>
sv -> join(sv, ", ")
end
2-element Vector{String}:
"hello, world"
"this, is, SPARTA"
(The ex.head == :vect is a basic sanity check to make sure that the string is in the format you expect, and not anything malicious, before evaluating it.)