Released under MIT license

nginx2csv.lua

-- For nginx combined log format:
-- http://nginx.org/en/docs/http/ngx_http_log_module.html#log_format
local query = '^'
  .. '(%d+%.%d+%.%d+%.%d+)%s' -- IP
  .. '%-%s' -- separator
  .. '(.-)%s' -- Remote user, TODO: Support
  .. '%[(.-)%]%s' -- Date, time
  .. '"(.-)"%s' -- request
  .. '(%d+)%s' -- HTTP status code
  .. '(%d+)%s' -- body_bytes_sent
  .. '"(.-)"%s' -- Referrer
  .. '"(.-)"' -- User agent
  -- Ignoring anything at the end of the string

for l in io.lines() do
  local ip,
        user,
        datetime,
        request,
        status,
        body_bytes_sent,
        referrer,
        user_agent
        = l:match(query)

  local request_method, request_url, request_protocol = "-", "-", "-"

  if ip and request ~= "-" then
    request_method, request_url, request_protocol = request:match(
        "^(%w+)%s(.-)%s(.-)$"
      )
  end

  if ip and request_method then
    io.write(
        ip, ",",
        user, ",",
        datetime, ",",
        request_method, ",",
        '"', request_url, '"', ",", -- should not contain '"'.
        request_protocol, ",",
        status, ",",
        body_bytes_sent, ",",
        '"', referrer, '"', ",", -- should not contain '"'.
        '"', user_agent, '"', ",", -- should not contain '"'.
        "\n"
      )
  else
    io.stderr:write("BAD:", l, "\n")
    io.stderr:flush()
  end
end