如果没有空格后跟大写字母，请删除空格

Question

I have data which looks like this:

*first*               *last*
M a rk                Twain
Hun ter               Stockt on Thompson

The data then continues for n amount of rows. So I want the data to look like this:

*first*               *last*
Mark                  Twain
Hunter                Stockton Thompson

I know I can use gsub to remove all blankspaces like this:

gsub(" ", "", x, fixed = TRUE)

And I can identify the pattern with a regex like this:

( [AZ])

But how can I combine these two to say to gsub - remove all spaces but not in the cases where it matches the regex?

Answer 1

Simplest way:

txt <- c("M a rk", "Twain", "Hun ter", "Stockt on Thompson")
gsub("\\s([a-z])", "\\1", txt)
## [1] "Mark"              "Twain"             "Hunter"            "Stockton Thompson"

If you want to apply this to more than one variable in a data.frame, you can do it using lapply and the list addressing replacement function for a data.frame. (Note: You really should not use asterisks in the names of data.frame columns.)

df <- data.frame("*first*" = c("M a rk", "Hun ter"),
                 "*last*" = c("Twain", "Stockt on Thompson"),
                 check.names = FALSE, stringsAsFactors = FALSE)

# names of the text columns you want to clean up
varsToModify <- c("*first*", "*last*")

df[varsToModify] <- lapply(df[varsToModify], 
                           function(x) gsub("\\s([a-z])", "\\1", x))
df
##   *first*            *last*
## 1    Mark             Twain
## 2  Hunter Stockton Thompson

Answer 2

df <- data.frame(`*first*`=c('M a rk','Hun ter'),`*last*`=c('Twain','Stockt on Thompson'),check.names=F,stringsAsFactors=F);
df;
##   *first*             *last*
## 1  M a rk              Twain
## 2 Hun ter Stockt on Thompson

I would use a Perl negative lookahead assertion:

for (ci in seq_along(df)) df[[ci]] <- gsub(perl=T,' (?![A-Z])','',df[[ci]]);
df;
##   *first*            *last*
## 1    Mark             Twain
## 2  Hunter Stockton Thompson

See Regular Expressions as used in R . The discussion of Perl assertions is given near the bottom of the page.