我有一些用户浏览行为的日志。它来自数据收集器，显然他用逗号分隔变量。但是有些网址里面有逗号。我无法将txt文件读入R。
20091,2009-06-02 22:06:14,84,taobao.com,search1.taobao.com,http://search1.taobao.com/browse/0/n-g,grdsa2kqn5scattbnzxq-------2-------b--40--commend-0-all-0.htm?at_topsearch=1&ssid=e-s1,www.taobao.com,shopping,e-commerce,C2C
20092,2009-06-16 12:25:35,8,sohu.com,www.wap.sohu.com,http://www.wap.sohu.com/info/index.html?url=http://wap.sohu.com/sports/pic/?lpn=1&resIdx=0&nid=336&rid=KL39,PD21746&v=2&ref=901981387,www.sohu.com,portal,entertainment,mobile
20092,2009-06-07 16:02:03,14,eetchina.com,www.powersystems.eetchina.com,http://www.powersystems.eetchina.com/ART_8800533274_2600005_TA_346f6b13.HTM?click_from=8800024853,8875136323,2009-05-26,PSCOL,ARTICLE_ALERT,,others,marketing,enterprise
20096,2009-06-30 07:51:38,7,taobao.com,search1.taobao.com,http://search1.taobao.com/browse/0/n-1----------------------0----------------------g,zhh3viy-g,ywtmf7glxeqnhjgt263ps-------2-------b--40--commend-0-all-0.htm?ssid=p1-s1,search1.taobao.com,shopping,e-commerce,C2C
2009184,2009-06-25 14:40:39,6,mktginc.com,surv.mktginc.com,,,unknown,unknown,unknown
20092,2009-06-07 15:13:06,32,ccb.com.cn,ibsbjstar.ccb.com.cn,https://ibsbjstar.ccb.com.cn/app/V5/CN/STY1/login.jsp,,e-bank,finance,e-bank
以上网址应为：
http://search1.taobao.com/browse/0/n-g,grdsa2kqn5scattbnzxq-------2-------b--40--commend-0-all-0.htm?at_topsearch=1&ssid=e-s1
http://www.wap.sohu.com/info/index.html?url=http://wap.sohu.com/sports/pic/?lpn=1&resIdx=0&nid=336&rid=KL39,PD21746&v=2&ref=901981387
http://www.powersystems.eetchina.com/ART_8800533274_2600005_TA_346f6b13.HTM?click_from=8800024853,8875136323,2009-05-26,PSCOL,ARTICLE_ALERT
http://search1.taobao.com/browse/0/n-1----------------------0----------------------g,zhh3viy-g,ywtmf7glxeqnhjgt263ps-------2-------b--40--commend-0-all-0.htm?ssid=p1-s1
https://ibsbjstar.ccb.com.cn/app/V5/CN/STY1/login.jsp
如何告诉R每行中确实有10个变量并在其中放入逗号？谢谢！
df <- read.table('2009.txt', sep= ',', quote= '', comment.char= '', stringsAsFactors= F)
Error in scan(file, what, nmax, sep, dec, quote, skip, nlines, na.strings, : line 130 did not have 10 elements
答案 0 :(得分：4)
你可以尝试：
dat <- read.table(text=gsub("http:.*(?=(,www)|,,)(*SKIP)(*F)|,", "*",
Lines, perl=TRUE), sep="*", header=FALSE, stringsAsFactors=FALSE)
dat
# V1 V2 V3 V4 V5
#1 20091 2009-06-02 22:06:14 84 taobao.com search1.taobao.com
#2 20092 2009-06-16 12:25:35 8 sohu.com www.wap.sohu.com
#3 20092 2009-06-07 16:02:03 14 eetchina.com www.powersystems.eetchina.com
# V6
#1 http://search1.taobao.com/browse/0/n-g,grdsa2kqn5scattbnzxq------- 2-------b--40--commend-0-all-0.htm?at_topsearch=1&ssid=e-s1
#2 http://www.wap.sohu.com/info/index.html?url=http://wap.sohu.com/sports/pic/?lpn=1&resIdx=0&nid=336&rid=KL39,PD21746&v=2&ref=901981387
#3 http://www.powersystems.eetchina.com/ART_8800533274_2600005_TA_346f6b13.HTM?click_from=8800024853,8875136323,2009-05-26,PSCOL,ARTICLE_ALERT
# V7 V8 V9 V10
#1 www.taobao.com shopping e-commerce C2C
#2 www.sohu.com portal entertainment mobile
#3 others marketing enterprise
Lines <- readLines(textConnection(txt)) #(`txt` from @Richard Scriven)
使用新数据集
indx <- grep("http", Lines)
Lines1 <- Lines[indx]
pat1 <- paste(unique(gsub(".*http[s]?.{3}(\\w+)\\..*", "\\1", Lines1)), collapse="|")
pat1N <- paste0("http:.*(?=,(", pat1, "|,))(*SKIP)(*F)|,")
dat1 <- read.table(text=gsub(pat1N, "*", Lines, perl=TRUE),
sep="*", header=FALSE, stringsAsFactors=FALSE)
dat1
# V1 V2 V3 V4 V5
#1 20091 2009-06-02 22:06:14 84 taobao.com search1.taobao.com
#2 20092 2009-06-16 12:25:35 8 sohu.com www.wap.sohu.com
#3 20092 2009-06-07 16:02:03 14 eetchina.com www.powersystems.eetchina.com
#4 20096 2009-06-30 07:51:38 7 taobao.com search1.taobao.com
#5 2009184 2009-06-25 14:40:39 6 mktginc.com surv.mktginc.com
#6 20092 2009-06-07 15:13:06 32 ccb.com.cn ibsbjstar.ccb.com.cn
# V6
# 1 http://search1.taobao.com/browse/0/n-g,grdsa2kqn5scattbnzxq-------2-------b--40--commend-0-all-0.htm?at_topsearch=1&ssid=e-s1
# 2 http://www.wap.sohu.com/info/index.html?url=http://wap.sohu.com/sports/pic/?lpn=1&resIdx=0&nid=336&rid=KL39,PD21746&v=2&ref=901981387
# 3 http://www.powersystems.eetchina.com/ART_8800533274_2600005_TA_346f6b13.HTM?click_from=8800024853,8875136323,2009-05-26,PSCOL,ARTICLE_ALERT
# 4 http://search1.taobao.com/browse/0/n-1----------------------0----------------------g,zhh3viy-g,ywtmf7glxeqnhjgt263ps-------2-------b--40--commend-0-all-0.htm?ssid=p1-s1
#5
#6 https://ibsbjstar.ccb.com.cn/app/V5/CN/STY1/login.jsp
# V7 V8 V9 V10
#1 www.taobao.com shopping e-commerce C2C
#2 www.sohu.com portal entertainment mobile
#3 others marketing enterprise
#4 search1.taobao.com shopping e-commerce C2C
#5 unknown unknown unknown
#6 e-bank finance e-bank
txt <- '20091,2009-06-02 22:06:14,84,taobao.com,search1.taobao.com,http://search1.taobao.com/browse/0/n-g,grdsa2kqn5scattbnzxq-------2-------b--40--commend-0-all-0.htm?at_topsearch=1&ssid=e-s1,www.taobao.com,shopping,e-commerce,C2C
20092,2009-06-16 12:25:35,8,sohu.com,www.wap.sohu.com,http://www.wap.sohu.com/info/index.html?url=http://wap.sohu.com/sports/pic/?lpn=1&resIdx=0&nid=336&rid=KL39,PD21746&v=2&ref=901981387,www.sohu.com,portal,entertainment,mobile
20092,2009-06-07 16:02:03,14,eetchina.com,www.powersystems.eetchina.com,http://www.powersystems.eetchina.com/ART_8800533274_2600005_TA_346f6b13.HTM?click_from=8800024853,8875136323,2009-05-26,PSCOL,ARTICLE_ALERT,,others,marketing,enterprise
20096,2009-06-30 07:51:38,7,taobao.com,search1.taobao.com,http://search1.taobao.com/browse/0/n-1----------------------0----------------------g,zhh3viy-g,ywtmf7glxeqnhjgt263ps-------2-------b--40--commend-0-all-0.htm?ssid=p1-s1,search1.taobao.com,shopping,e-commerce,C2C
2009184,2009-06-25 14:40:39,6,mktginc.com,surv.mktginc.com,,,unknown,unknown,unknown
20092,2009-06-07 15:13:06,32,ccb.com.cn,ibsbjstar.ccb.com.cn,https://ibsbjstar.ccb.com.cn/app/V5/CN/STY1/login.jsp,,e-bank,finance,e-bank'
Lines <- readLines(textConnection(txt))
答案 1 :(得分：3)
如果你读入数据使得每一行都是一个字符串（例如sep="\n"），那么你可以在放入正确的数据框之前直接处理每一行。
如果只有第6个条目可能有逗号（看起来其他网址只是主要网域），那么以下内容可能会有效：
d <- strsplit(d, ",")
for (i in 1:length(d)) {
x <- d[[i]]
n <- length(x)
if (n > 10) {
d[[i]] <- c(x[1:5], paste(x[6:(n-4)], collapse=","), x[(n-3):n])
}
}
d <- do.call(rbind,lapply(d, matrix, ncol=10, byrow=TRUE))
如果其他网址可能存在问题，这种方法可能仍然有效，但可能会变得非常复杂。
答案 2 :(得分：1)
这看起来可能对你有用。
txt <- '20091,2009-06-02 22:06:14,84,taobao.com,search1.taobao.com,http://search1.taobao.com/browse/0/n-g,grdsa2kqn5scattbnzxq-------2-------b--40--commend-0-all-0.htm?at_topsearch=1&ssid=e-s1,www.taobao.com,shopping,e-commerce,C2C
20092,2009-06-16 12:25:35,8,sohu.com,www.wap.sohu.com,http://www.wap.sohu.com/info/index.html?url=http://wap.sohu.com/sports/pic/?lpn=1&resIdx=0&nid=336&rid=KL39,PD21746&v=2&ref=901981387,www.sohu.com,portal,entertainment,mobile
20092,2009-06-07 16:02:03,14,eetchina.com,www.powersystems.eetchina.com,http://www.powersystems.eetchina.com/ART_8800533274_2600005_TA_346f6b13.HTM?click_from=8800024853,8875136323,2009-05-26,PSCOL,ARTICLE_ALERT,,others,marketing,enterprise'
readLog <- function(file, stringsAsFactors = TRUE)
{
s <- strsplit(readLines(file), ",")
loop <- t(sapply(s, function(x) {
g <- grep("http", x)
x[g] <- paste(x[g:(length(x)-4)], collapse = ",")
x[-c((g+1):(length(x)-4))]
}))
data.frame(loop, stringsAsFactors = stringsAsFactors)
}
## readLog(textConnection(txt))
readLog(yourFile)
这在第6列中给出以下内容，每行有10列
V6
1 http://search1.taobao.com/browse/0/n-g,grdsa2kqn5scattbnzxq-------2-------b--40--commend-0-all-0.htm?at_topsearch=1&ssid=e-s1
2 http://www.wap.sohu.com/info/index.html?url=http://wap.sohu.com/sports/pic/?lpn=1&resIdx=0&nid=336&rid=KL39,PD21746&v=2&ref=901981387
3 http://www.powersystems.eetchina.com/ART_8800533274_2600005_TA_346f6b13.HTM?click_from=8800024853,8875136323,2009-05-26,PSCOL,ARTICLE_ALERT
7到10是
V7 V8 V9 V10
1 www.taobao.com shopping e-commerce C2C
2 www.sohu.com portal entertainment mobile
3 others marketing enterprise