R help - Oct 2017 - Select part of character row name in a data frame

Dear R contributors,

I have a problem in selecting in an efficient way, rows of a data frame
according to a condition,
which is a part of a row name of the table.

The data frame is made of 64 rows and 2 columns, but the row names are very long
but I need to select them according to a small part of it and perform
calculations on the subsets.

This is the example:
											X			Y
"Unique to strat          "                                           
0.0482        28.39
"Unique to crt.dummy        "                                         
0.0441        25.92
"Unique to gender                   "                                 
0.0159         9.36
"Unique to age                       "                                
0.0839        49.37
"Unique to gg_right1              "                                   
0.0019         1.10
"Unique to strat:crt.dummy                 "                          
0.0689        40.54
"Common to strat, and crt.dummy         "                             
-0.0392       -23.09
"Common to strat, and gender         "                                
-0.0031        -1.84
"Common to crt.dummy, and gender     "                                
0.0038         2.21
"Common to strat, and age                 "                           
0.0072         4.21

X and Y are the two columns of variables, while ?Unique to strat?, are the row
names. I am interested to select for example those rows
whose name contains ?strat? only. It would be very easy if these names were
simple, but they are not and involve also spaces.
I tried with select matches from dplyr but works for column names but I did not
find how to use it on row names, which are of course character values.

Thanks for any help you can provide.

----------------------------------
Francesca Pancotto, PhD

Quoting Francesca PANCOTTO <f.pancotto at unimore.it>:
> Dear R contributors,
>
> I have a problem in selecting in an efficient way, rows of a data  
> frame according to a condition,
> which is a part of a row name of the table.
>
> The data frame is made of 64 rows and 2 columns, but the row names  
> are very long but I need to select them according to a small part of  
> it and perform calculations on the subsets.
>
> This is the example:
> 											X			Y
> "Unique to strat          "
>           0.0482        28.39
> "Unique to crt.dummy        "
>           0.0441        25.92
> "Unique to gender                   "
>           0.0159         9.36
> "Unique to age                       "
>           0.0839        49.37
> "Unique to gg_right1              "
>           0.0019         1.10
> "Unique to strat:crt.dummy                 "
>           0.0689        40.54
> "Common to strat, and crt.dummy         "
>          -0.0392       -23.09
> "Common to strat, and gender         "
>          -0.0031        -1.84
> "Common to crt.dummy, and gender     "
>           0.0038         2.21
> "Common to strat, and age                 "
>           0.0072         4.21
>
> X and Y are the two columns of variables, while ?Unique to strat?,  
> are the row names. I am interested to select for example those rows
> whose name contains ?strat? only. It would be very easy if these  
> names were simple, but they are not and involve also spaces.
> I tried with select matches from dplyr but works for column names  
> but I did not find how to use it on row names, which are of course  
> character values.
>
> Thanks for any help you can provide.
>
> ----------------------------------
> Francesca Pancotto, PhD
>
Use ?grep or ?grepl:

     df[grep("strat", row.names(df)), ]

(in which 'df' is your data frame)


-- 
Enrico Schumann
Lucerne, Switzerland
http://enricoschumann.net

Thanks a lot, so simple so efficient!

I will study more the grep command I did not know.

Thanks!


Francesca Pancotto
> Il giorno 19 ott 2017, alle ore 12:12, Enrico Schumann <es at
enricoschumann.net> ha scritto:
> 
>  df[grep("strat", row.names(df)), ]

	[[alternative HTML version deleted]]