Difference between contains() and matches() for se

问题:

I have decided to spend some time to learn dplyr thoroughly. I have just come across the select() function and some of the helper functions that come with it.

By just playing around I have failed to find any difference between the contains and matches helper functions.

Could someone please provide an example of how they can be used for different purposes?

Thank you,

回答1:

The difference is that matches can take regex as pattern to match column names and select while contains does the literal match of substring or full name match. It is described in the ?select_helpers as

contains(): Contains a literal string.

matches(): Matches a regular expression.

Consider a simple example where we want to select columns that have substring 'col'

df1 <- data.frame(colnm = 1:5, col1 = 24, col2 = 46)
df1 %>% 
    select(contains("col"))
#  colnm col1 col2
#1     1   24   46
#2     2   24   46
#3     3   24   46
#4     4   24   46
#5     5   24   46

Here, it matches the 'col' literally in the column names and select those. If we change the matching criteria to match 'col' followed by one or more digits (\\d+) with a regex

df1 %>% 
   select(contains("col\\d+"))
#data frame with 0 columns and 5 rows

if fails, because it is looking for column name substring "col\\d+"

df1 %>%
    select(matches("col\\d+")) 
# col1 col2
#1   24   46
#2   24   46
#3   24   46
#4   24   46
#5   24   46

whereas matches take regex and match those patterns