用R中的Rex包解析服务器日志

Question

我有要解析的服务器日志数据格式。

在这里前两行

test <- c("5638052581 \"Norway|Oslo County|Oslo|3163036322|503858711|160449504|y|\" n - - [31/Oct/2019:13:00:01 +0000] \"GET /P04_AL?args=app_01&distributor=p4&player=app&playeros=ios&referrer=1&station=1&codec=aac&quality=low&deviceid=1D6A84DA-92A6-4AD1-A2A3-1AB20D2263B2&listenerid=61D1F2EB-7B35-4434-9D8B-A6D074BE28F0&userid=fczUdjf5yEU8j4JlZHG4JXABgiZ2&aw_0_1st.audience=%5B%22P7ActiveListeners%22,%20%22p5hitsactive%22,%20%22P6ActiveListeners%22,%20%22P4ActiveListeners%22,%20%22AppInstalledP4%22%5D HTTP/1.1\" 200 4305805 \"-\" \"AppleCoreMedia//1.0.0.17B84 (iPhone; U; CPU OS 13_2 like Mac OS X; nb_no)\" 702", "616118387 \"Netherlands|North Holland|Haarlem|631068861|616118387|862817723||\" n - - [31/Oct/2019:13:00:01 +0000] \"GET /P04_MH HTTP/1.1\" 200 519546 \"-\" \"MultiRoomAudioPlayer//5.1\" 6")

我试图像下面这样使用雷克斯包，但是经常会遇到意外输入的错误。我做错什么了？有人能帮我做这个吗。以下是我对一个记录的尝试(向量的第一个元素)

library(rex)
re_logic <- rex(
  
  capture(name = "process_id", digits),
  "`\´",
  capture(name = "country", non_spaces),
  "|",
  capture(name = "county", non_spaces),
  "|",
  capture(name = "city", non_spaces), 
  "|",
  capture(name = "x1", digits), 
  "|",
  capture(name = "x2", digits),
  "|",
  capture(name = "x3", digits),
  "|",
  capture(name = "process_name", alpha),
  "`n - -´",
  spaces,
  "[",
  capture(name = "accept_date", except_some_of("]")),
  "]",
  spaces,
  "`\´",
  capture(name = "http_request", non_quotes),
  "`\´",
  spaces,
  capture(name = "status_code", digits),
  spaces,
  capture(name = "bytes_read", some_of("+", digit)),
  "`" \"´",
  capture(name = "actconn", digits),
  "`//´",
  spaces,
  "(",
  capture(name = "Tr", non_quotes),
  ";" )
  

# sample view
re_matches(test, re_logic) %>% as_tibble()

Answer 1

您可以使用

re_logic <- rex(
  capture(name = "process_id", digits),
  spaces, quote,
  capture(name = "country", except_some_of("|")),
  "|",
  capture(name = "county", except_some_of("|")),
  "|",
  capture(name = "city", except_some_of("|")), 
  "|",
  capture(name = "x1", digits), 
  "|",
  capture(name = "x2", digits),
  "|",
  capture(name = "x3", digits),
  "|",
  capture(name = "process_name",  zero_or_more(alpha)),
  "|", quote, spaces, "n", spaces, "-", spaces, "-",spaces,
  "[",
  capture(name = "accept_date", except_some_of("]","[")),
  "]",
  spaces, quote,
  capture(name = "http_request", non_quotes),
  quote, spaces,
  capture(name = "status_code", digits),
  spaces,
  capture(name = "bytes_read", some_of("+", digit)),
  spaces, quote, non_quotes, quote, spaces, quote,
  capture(name = "actconn", except_some_of(quote, "/")),
  "/", non_spaces,
  maybe(
    spaces, "(",
    capture(name = "Tr", except_some_of(";"))
  )
)
re_matches(test, re_logic)

见regex演示。

注意到

• 我使用quote来匹配任何'或"字符
• 我没有使用non_spaces来匹配地理名称，而是使用了任何字符(除了|模式，except_some_of("|") )
• Tr部件是可选的，因此需要用maybe子句包装与该组相关的模式链。