How to assign output of cat to an object?

Question

How would it be possible in the example below to skip the step of writing to file "test.txt", ie assign the cat-result to an object, and still achieve the same end result?

I thought I'd include the full example to give background to my problem.

test <- c("V 1", "x", "1 2 3", "y", "3 5 8", "V 2", "x", "y", "V 3", "y", "7 2 1", "V 4", "x", "9 3 7", "y")

# Write selection to file
cat(test, "\n", file="test.txt")
test2 <- readLines("test.txt")
test3 <- strsplit(test2, "V ")[[1]][-1]

# Find results
x <- gsub("([0-9]) (?:x )?([0-9] [0-9] [0-9])?.*", "\\1 \\2 ", test3, perl = TRUE)
y <- gsub("([0-9]).* y ?([0-9] [0-9] [0-9])?.*", "\\1 \\2 ", test3, perl = TRUE)

# Eliminate tests with no results
x1 <- x[regexpr("[0-9] ([^0-9]).*", x) == -1]
y1 <- y[regexpr("[0-9] ([^0-9]).*", y) == -1]

# Dataframe of results
xdf1 <- read.table(textConnection(x1), col.names=c("id","x1","x2","x3"))
ydf1 <- read.table(textConnection(y1), col.names=c("id","y1","y2","y3"))
closeAllConnections()

# Dataframe of tests with no results
x2 <- x[regexpr("[0-9] ([^0-9]).*", x) == 1]
y2 <- y[regexpr("[0-9] ([^0-9]).*", y) == 1]

df1 <- as.integer(x2[x2 == y2])
df1 <- data.frame(id = df1)

# Merge dataframes
results <- merge(xdf1, ydf1, all = TRUE)
results <- merge(results, df1, all = TRUE)
results

Results in:

  id x1 x2 x3 y1 y2 y3
1  1  1  2  3  3  5  8
2  2 NA NA NA NA NA NA
3  3 NA NA NA  7  2  1
4  4  9  3  7 NA NA NA

Answer 1

As a more general solution, you can use the capture output function. It results in a character vector with elements corresponding to each line of the output.

your example:

test2<-capture.output(cat(test))

here is a multi-line example:

> out<-capture.output(summary(lm(hwy~cyl*drv,data=mpg)))
> out
 [1] ""                                                               
 [2] "Call:"                                                          
 [3] "lm(formula = hwy ~ cyl * drv, data = mpg)"                      
 [4] ""                                                               
 [5] "Residuals:"                                                     
 [6] "    Min      1Q  Median      3Q     Max "                       
 [7] "-8.3315 -1.4139 -0.1382  1.6479 13.5861 "                       
 [8] ""                                                               
 [9] "Coefficients:"                                                  
[10] "            Estimate Std. Error t value Pr(>|t|)    "           
[11] "(Intercept)  32.1776     1.2410  25.930  < 2e-16 ***"           
[12] "cyl          -2.0049     0.1859 -10.788  < 2e-16 ***"           
[13] "drvf          8.4009     1.8965   4.430 1.47e-05 ***"           
[14] "drvr          8.2509     6.4243   1.284    0.200    "           
[15] "cyl:drvf     -0.5362     0.3422  -1.567    0.119    "           
[16] "cyl:drvr     -0.5248     0.8379  -0.626    0.532    "           
[17] "---"                                                            
[18] "Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 "
[19] ""                                                               
[20] "Residual standard error: 2.995 on 228 degrees of freedom"       
[21] "Multiple R-squared: 0.7524,\tAdjusted R-squared: 0.747 "         
[22] "F-statistic: 138.6 on 5 and 228 DF,  p-value: < 2.2e-16 "       
[23] ""    

Answer 2

Instead of cat ing to a file, why not use the paste command to generate a string instead?

> paste(test, collapse="\n")
[1] "V 1\nx\n1 2 3\ny\n3 5 8\nV 2\nx\ny\nV 3\ny\n7 2 1\nV 4\nx\n9 3 7\ny"

Now instead of doing a cat then readlines you can just pass this string directly into strsplit .

Answer 3

Try

> f <- textConnection("test3", "w")
> cat(test, "\n", file=f)
> test3
[1] "V 1 x 1 2 3 y 3 5 8 V 2 x y V 3 y 7 2 1 V 4 x 9 3 7 y "
> close(f)

Answer 4

There's also the assign statement which allows you to build a name and set an object to it. Very useful if you want to iterate a bunch of tests and name them with dynamic values.

assign("mary", paste(test,sep= "\\n"))

will assign the paste statement to mary. However say you were running a bunch of regressions and wanted your regression objects named by predictor. You could do something like

assign(paste("myRegression",names(dataframe)[2],sep=""), lm(dataframe$response~dataframe[,2]))

which would give you the object

myRegressionPredictorName as you linear model.

Answer 5

Try the following codes:

writeLines(capture.out((summary(lm(hwy~cyl*drv,data=mpg)),con="summary.txt",sep="\\n")

Then you can open the txt file "summary.txt" to see your results.