What is the command for printing the structure of R objects so the object can be re-created by running the printed output?
The output often contains the structure function, and you can copy and paste the output into your code in order to easily create an object for a reproducible example.
I'm breaking my head all morning over this command that I ought to know.
The function is dput (or dump).
I think str function (aka structure) has more consistently useful information than dump.
From the R help:
Compactly Display the Structure of an Arbitrary R Object
Description
Compactly display the internal structure of an R object, a diagnostic function and an
alternative to summary (and to some extent, dput). Ideally, only one line for each
‘basic’ structure is displayed. It is especially well suited to compactly display the
(abbreviated) contents of (possibly nested) lists. The idea is to give reasonable
output for any R object. It calls args for (non-primitive) function objects.
Related
I have a list whose elements are several dataframes, which looks like this
Because it is hard for another user to use these data by re-running my original code. Hence, I would like to export it. As the graph shows, the dataframes in that list have different number of rows. I am wondering if there is any method to export it as file without damaging any information, and make it be able to be used by Rstudio. I have tried to save it as RData, but I don't know how to save the information.
Thanks a lot
To output objects in R, here are 4 common methods:
dput() writes a text representation of an R object
This is very convenient if you want to allow someone to get your object by copying and pasting text (for instance on this site), without having to email or upload and download a file. The downside however is that the output is long and re-reading the object into R (simply by assigning the copied text to an object) can hang R for large objects. This works best to create reproducible examples. For a list of data frames, this would not be a very good option.
You can print an object to a .csv, .xlsx, etc. file with write.table(), write.csv(), readr::write_csv(), xlsx::write.xlsx(), etc.
While the file can then be used by other software (and re-imported into R with read.csv(), readr::read_csv(), readxl::read_excel(), etc.), the data can be transformed in the process and some objects cannot be printed in a single file without prior modifications. So this is not ideal in your case either.
save.image() saves your entire workspace (objects + environment)
The workspace can then be recreated with load(). This can be useful, but you are here only interested in saving one object. In that case, it is preferable to use:
saveRDS() which allows to write one object to file
The object can then be re-created with readRDS(). This is the best option to save an R object to file, without any modification and then re-create it.
In your situation, this is definitely the best solution.
I have a dataframe loaded successfully in R.
I would like to give the data of df to someone else to use them with quick and easy way without need to load again the file into a df.
Which is the command to give the whole data of df (not the str())
You can save the file into a .RData using save or save.image, depending on your needs. First one will save specific objects while the latter will dump the whole workspace to a file. This method has the advantage of working on probably any R object.
Another option is as #user1945827 mentioned, using dput which will produce a string that is parseable into another R session. This will not work for complex (like S4) objects.
I have multiple regressions in an R script and want to append the regression summaries to a single text file output. I know I can use the following code to do this for one regression summary, but how would I do this for multiple?
rpt1 <- summary(fit)
capture.output(rpt1, file = "results.txt")
I would prefer not to have to use this multiple times in the same script (for rpt1, rpt2, etc.), and thus have separate text files for each result. I'm sure this is easy, but I'm still learning the R ropes. Any ideas?
You can store the result as a list and then use the capture.output
fit1<-lm(mpg~cyl,data=mtcars)
fit2<-lm(mpg~cyl+disp,data=mtcars)
myresult<-list(fit1,fit2)
capture.output(myresult, file = "results.txt")
If you want multiple output sent to a file then look at the sink function, it will redirect all output to a file until you call sink again. The capture.output function actually uses sink.
You might also be interested in the txtStart function (and friends) in the TeachingDemos package which will also include the commands interspersed with the output and gives a few more options for output formatting.
Eventually you will probably want to investigate the knitr package for ways of running a set of commands in a batch and nicely capturing all the output together nicely formatted (and documented).
I'm compiling knitr document from a .R file:
knit2pdf(example.Rnw, output=paste0(name,".tex"))
But in the document example.Rnw, there's the function:
do.call(cbind,mget(as.character(rep_names)))
Where rep_names it's the name of some dataframes created by a loop, which i want to merge. The thing is i don't know the extact number of dataframes created.
If i compile the document directly from knitr works perfectly, but when i execute from .R file, he didn't find the objects of rep_names. Example on .pdf output.
## Error: value for ’Object_1’ not found
Where Object_1 is rep_names[1]. The problem is on which session are the object created?
It's hard to provide any assistance here without a reproducible example. Looking at what you have, I imagine you should change the way you're creating the dataframes to put them in a list rather than as named objects in the global environment. This way you can change the problematic line to do.call(cbind, listofdfs). There are numerous questions and answers here that recommend this strategy.
I have multiple regressions in an R script and want to append the regression summaries to a single text file output. I know I can use the following code to do this for one regression summary, but how would I do this for multiple?
rpt1 <- summary(fit)
capture.output(rpt1, file = "results.txt")
I would prefer not to have to use this multiple times in the same script (for rpt1, rpt2, etc.), and thus have separate text files for each result. I'm sure this is easy, but I'm still learning the R ropes. Any ideas?
You can store the result as a list and then use the capture.output
fit1<-lm(mpg~cyl,data=mtcars)
fit2<-lm(mpg~cyl+disp,data=mtcars)
myresult<-list(fit1,fit2)
capture.output(myresult, file = "results.txt")
If you want multiple output sent to a file then look at the sink function, it will redirect all output to a file until you call sink again. The capture.output function actually uses sink.
You might also be interested in the txtStart function (and friends) in the TeachingDemos package which will also include the commands interspersed with the output and gives a few more options for output formatting.
Eventually you will probably want to investigate the knitr package for ways of running a set of commands in a batch and nicely capturing all the output together nicely formatted (and documented).