How to extract data from a text file using R or Po

问题:

I have a text file containing data like this:

This is just text
-------------------------------
Username:          SOMETHI           C:                 [Text]
Account:           DFAG              Finish time:        1-JAN-2011 00:31:58.91
Process ID:        2028aaB           Start time:        31-DEC-2010 20:27:15.30

This is just text
-------------------------------
Username:          SOMEGG            C:                 [Text]
Account:           DFAG              Finish time:        1-JAN-2011 00:31:58.91
Process ID:        20dd33DB          Start time:        12-DEC-2010 20:27:15.30

This is just text
-------------------------------
Username:          SOMEYY            C:                 [Text]
Account:           DFAG              Finish time:        1-JAN-2011 00:31:58.91
Process ID:        202223DB          Start time:        15-DEC-2010 20:27:15.30

Is there a way to extract Username, Finish time, Start time from this kind of data? I'm looking for some starting point usign R or Powershell.

回答1:

R may not be the best tool to process text files, but you can proceed as follows: identify the two columns by reading the file as a fixed-width file, separate the fields from their value by splitting the strings on the colons, add an "id" column, and put everything back in order.

# Read the file
d <- read.fwf("A.txt", c(37,100), stringsAsFactors=FALSE)

# Separate fields and values
d <- d[grep(":", d$V1),]
d <- cbind( 
  do.call( rbind, strsplit(d$V1, ":\\s+") ), 
  do.call( rbind, strsplit(d$V2, ":\\s+") ) 
)

# Add an id column
d <- cbind( d, cumsum( d[,1] == "Username" ) )

# Stack the left and right parts
d <- rbind( d[,c(5,1,2)], d[,c(5,3,4)] )
colnames(d) <- c("id", "field", "value")
d <- as.data.frame(d)
d$value <- gsub("\\s+$", "", d$value)

# Convert to a wide data.frame
library(reshape2)
d <- dcast( d, id ~ field )

回答2:

These are just guidelines of how I would approach the problem. I'm sure there's a more fancy way of doing it. Possibly including plyr. :)

rara <- readLines("test.txt") # you could use readLines(textConnection = "text"))

# find usernames
usn <- rara[grepl("Username:", rara)]
# you can find a fancy way to split or weed out spaces
# I crudely do it like this:
unlist(lapply(strsplit(usn, "      "), "[", 2)) # 2 means "extract the second element"

# and accounts
acc <- rara[grepl("Account:", rara)]
unlist(lapply(strsplit(acc, "      "), "[", 2))

You can use str_trim() to remove whitespace before/after the word. Hope there's enough pointers to get you going.

回答3:

Here's a Powershell solution:

$result = @()

get-content c:\somedir\somefile.txt |
foreach {
    if ($_ -match '^Username:\s+(\S+)'){
        $rec = ""|select UserName,FinishTime,StartTime
        $rec.UserName = $matches[1]
        }
    elseif ($_ -match '^Account.+Finish\stime:\s+(.+)'){
        $rec.FinishTime = $matches[1]
        }
    elseif ($_ -match '^Process\sID:\s+\S+\s+Start\stime:\s+(.+)'){
        $rec.StartTime = $matches[1]
        $result += $rec
        }
}
$result

回答4:

Do you have your file in a data frame? Like the column names would be Username, Process ID, Start time... If so, you can easly extract it by

df$Username (where df is your data frame and if you want to see all your usernames)
df$FinishTime

If you want to know everything about a user with a certain name, use this

df[df$username == "SOMETHI",]

If you want to know a user with a finish time..

Hope this can be a starting point. Let me know if sth is not clear.