Include empty matches when using string.gmatch to split a string in lua 5.1

Question

I have a comma seperated input string that needs to support empty entries. So a string like a,b,c,,d should result in a table with 5 entries, where the 4'th is an empty value.

A simplified example

str="a,b,c,,d"
count=0

for v in string.gmatch(str, '([^,]*)') do
    count = count + 1
end

print(count)

This code outputs

9

in Lua 5.1, although there are only 5 entries.

I can change the * in the regex to + - then it reports the 4 entries a,b,c,d but not the empty one. It seems that this behaviour has been fixed in Lua 5.2, because the code above works fine in lua 5.2, but I'm forced to find a solution for lua 5.1

My current implementation

function getValues(inputString)
  local result = {}

  for v in string.gmatch(inputString, '([^,]*)') do
    table.insert(result, v)
  end

  return result
end

Any suggestions about how to fix?

simple workaround: replace the empty positions with a space or "novalue" or whatever suits your needs. — Piglet, Apr 06 '20 at 07:37

score 3 · Accepted Answer · answered Apr 06 '20 at 09:50

3

You may append a comma to the text and grab all values using a ([^,]*), pattern:

function getValues(inputString)
  local result = {}

  for v in string.gmatch(inputString..",", '([^,]*),') do
    table.insert(result, v)
  end

  return result
end

The output:

a
b
c

d

answered Apr 06 '20 at 09:50

Wiktor Stribiżew

607,720
39
448
563

@lhf Negated character classes usually perform better. – Wiktor Stribiżew Apr 06 '20 at 11:36

DarkWiiPlayer · Answer 2 · 2020-04-06T08:01:45.537

0

local str="a,b,c,,d"
local count=1
for value in string.gmatch(str, ',') do
    count = count + 1
end
print(count)

And if you want to get the values, you could do something like


local function values(str, previous)
    previous = previous or 1
    if previous <= #str then
        local comma = str:find(",", previous) or #str+1
        return str:sub(previous, comma-1), values(str, comma+1)
    end
end

edited Apr 06 '20 at 08:01

answered Apr 06 '20 at 07:52

DarkWiiPlayer

6,871
3
23
38

Sorry, it wasn't clear in my question that I need the values. I have updated the question accordingly. Would you mind providing an example of how to use this `values` function? – Daniel Apr 06 '20 at 08:47
You can just do `v = {values("a,b,c,,d")}` to get a table with all the values :D – DarkWiiPlayer Apr 06 '20 at 09:49

Include empty matches when using string.gmatch to split a string in lua 5.1

My current implementation

2 Answers2