解析错误：尝试解析数据框中的 JSON 列时出现“尾随垃圾”

Question

我有一个看起来像 this 的日志文件。这是一个文本文档，如下所示：

Id,Date,Level,Message
35054,2016-06-17 19:29:43 +0000,INFO,"{
  ""id"": -2,
  ""ipAddress"": ""100.100.100.100"",
  ""howYouHearAboutUs"": null,
  ""isInterestedInOffer"": true,
  ""incomeRange"": 60000,
  ""isEmailConfirmed"": false
}"
35055,2016-06-17 19:36:38 +0000,INFO,"{
  ""id"": -1,
  ""firstName"": ""John"",
  ""lastName"": ""Smith"",
  ""email"": ""john.smith@gmail.com"",
  ""city"": ""Smalltown"",
  ""incomeRange"": 1,
  ""birthDate"": ""1999-12-10T05:00:00Z"",
  ""password"": ""*********"",
  ""agreeToTermsOfUse"": true,
  ""howYouHearAboutUs"": ""Radio"",
  ""isInterestedInOffer"": false
}"
35059,2016-07-19 19:52:08 +0000,INFO,"{
  ""id"": -3,
  ""visitUrl"": ""https://www.website.com/?purpose=X"",
  ""ipAddress"": ""100.200.300.400"",
  ""howYouHearAboutUs"": null,
  ""isInterestedInOffer"": true,
  ""incomeRange"": 100000,
  ""isEmailConfirmed"": true,
  ""isIdentityConfirmed"": false,
  ""agreeToTermsOfUse"": true,
  ""validationResults"": null
}"

我正在尝试通过以下方式解析 Message 列中的 JSON：

library(readr)
library(jsonlite)

df <- read_csv("log_file_from_above.csv")
fromJSON(as.character(df$Message))

但是，我遇到了以下错误：

Error: parse error: trailing garbage
          "isEmailConfirmed": false  } {    "id": -1,    "firstName": 
                     (right here) ------^

我怎样才能摆脱“尾随垃圾”？

Answer 1

fromJSON() 不是针对字符向量“应用”，而是试图将其全部转换为数据帧。你可以试试

purrr::map(df$Message, jsonlite::fromJSON)

@Abdou 提供了什么或

jsonlite::stream_in(textConnection(gsub("\\n", "", df$Message)))

后两者将创建数据框。第一个将创建一个列表，您可以将其添加为列。

您可以使用dplyr::bind_cols 的最后一种方法来创建一个包含所有数据的新数据框：

dplyr::bind_cols(df[,1:3],
                 jsonlite::stream_in(textConnection(gsub("\\n", "", df$Message))))

@Abdou 还建议使用几乎纯 R 的基础解决方案：

cbind(df, do.call(plyr::rbind.fill, lapply(paste0("[",df$Message,"]"), function(x) jsonlite::fromJSON(x))))

完整、有效的工作流程：

library(dplyr)
library(jsonlite)

df <- read.table("http://pastebin.com/raw/MMPMwNZv",
                 quote='"', sep=",", stringsAsFactors=FALSE, header=TRUE)

bind_cols(df[,1:3], stream_in(textConnection(gsub("\\n", "", df$Message)))) %>%
  glimpse()
## 
 Found 3 records...
 Imported 3 records. Simplifying into dataframe...
## Observations: 3
## Variables: 19
## $ Id                  <int> 35054, 35055, 35059
## $ Date                <chr> "2016-06-17 19:29:43 +0000", "2016-06-17 1...
## $ Level               <chr> "INFO", "INFO", "INFO"
## $ id                  <int> -2, -1, -3
## $ ipAddress           <chr> "100.100.100.100", NA, "100.200.300.400"
## $ howYouHearAboutUs   <chr> NA, "Radio", NA
## $ isInterestedInOffer <lgl> TRUE, FALSE, TRUE
## $ incomeRange         <int> 60000, 1, 100000
## $ isEmailConfirmed    <lgl> FALSE, NA, TRUE
## $ firstName           <chr> NA, "John", NA
## $ lastName            <chr> NA, "Smith", NA
## $ email               <chr> NA, "john.smith@gmail.com", NA
## $ city                <chr> NA, "Smalltown", NA
## $ birthDate           <chr> NA, "1999-12-10T05:00:00Z", NA
## $ password            <chr> NA, "*********", NA
## $ agreeToTermsOfUse   <lgl> NA, TRUE, TRUE
## $ visitUrl            <chr> NA, NA, "https://www.website.com/?purpose=X"
## $ isIdentityConfirmed <lgl> NA, NA, FALSE
## $ validationResults   <lgl> NA, NA, NA