How to split some rows into columns in R?

Question

If I have a data table in R like this:

Table0:

some subtitle 1
info1a | info2a / info3a / info4a
info1b | info2b / info3b / info4b
some subtitle 2
info1c | info2c / info3c / info4c
info1d | info2d / info3d / info4d
...

Is there a way to delete all the subtitle rows and then split the info columns normally (by | and /)? The subtitles can ideally go in a separate table.

Thus, the result would be:

Table1, where the pipes represent the breaks between the columns:

info1a | info2a | info3a | info4a
info1b | info2b | info3b | info4b
info1c | info2c | info3c | info4c
info1d | info2d | info3d | info4d
...

Table2:

some subtitle 1
some subtitle 2
...

I don't understand your tables. What are the columns in Table0 (is it a `data.frame` with a single column, or maybe a vector?). What is Table2? Is it a `data.frame` with a single column? — jbaums, Jan 02 '16 at 04:16
Use a combination of `split` and `strsplit`. Post some more meaningful sample data for better suggestions..... — A5C1D2H2I1M1N2O1R2T1, Jan 02 '16 at 04:17
This question can easily be salvaged with some sample data. Joe: Please consider doing so. — A5C1D2H2I1M1N2O1R2T1, Jan 02 '16 at 05:05
It's not clear, for example, what the structure of "Table0" is. Is it a single-column `data.frame`? A two-column `data.frame`? — A5C1D2H2I1M1N2O1R2T1, Jan 02 '16 at 05:07
It looks like a combination of splitting a column into multiples and indexing internal headers. The question is definitely salvageable. — Pierre L, Jan 02 '16 at 05:21
Table0 can just be a .txt file, and it can be loaded into R however is most convenient. @AnandaMahto do you have any suggestions for sample data? — John Targaryen, Jan 03 '16 at 17:05
@JoeBob, The best would be if you could share a few lines from one of your text files. It's not clear, for example, how one should identify the "some subtitle" lines. (I'm guessing that the lines don't actually read "some subtitle"). — A5C1D2H2I1M1N2O1R2T1, Jan 03 '16 at 17:16
Similarly, it's important for us to know whether your `|` is actually a pipe character, and your `/` an actual slash character, or whether you are just showing those as an example. — A5C1D2H2I1M1N2O1R2T1, Jan 03 '16 at 17:17

score 3 · Accepted Answer · edited May 23 '17 at 12:23

3

Slicing

rowsiwant <- c(1,4)
#For the first table by removing the unwanted rows
new_table_data <- old_table[-rowsiwant,]
#Creates the second table 
new_table_labels <- old_table[rowsiwant,]

Or you can try the subset function if your row names are numbers:

#Creates the first table
new_table_labels <- subset(old_table, rownames(old_table) != rowsiwant)
#Creates the second table
new_table_data <- subset(old_table, rownames(old_table) == rowsiwant)

Also, there are plenty of great resources for slicing such as: http://statmethods.net/management/subset.html http://statistics.ats.ucla.edu/stat/r/faq/subset_R.htm

Also previous questions are a good resource:

How can I subset rows in a data frame in R based on a vector of values?

Subset of table in R using row numbers?

edited May 23 '17 at 12:23

Community

1
1

answered Jan 02 '16 at 05:23

Cody Glickman

514
1
8
30

What if I wanted to cut out more rows? Like 1, 4, 7, 10 ... 1003, how would I remove those rows? Would I have to type in each number, or is there a simpler way? – John Targaryen Jan 03 '16 at 16:50
You could add each row individually, though if you have a pattern (say every third row) of labels you can slice accordingly. I would recommend a grep for rows with subtitles (if they have the same subtitle) http://stackoverflow.com/questions/21311386/using-grep-to-help-subset-a-data-frame-in-r – Cody Glickman Jan 04 '16 at 20:38
Also, http://stackoverflow.com/questions/19198492/subset-data-frame-by-complex-pattern-of-column-names & http://stackoverflow.com/questions/19417674/subset-with-pattern are good examples of regex – Cody Glickman Jan 04 '16 at 20:42

How to split some rows into columns in R?

1 Answers1