R：将长单列dataframe展开成两列，按字母和数字拆分，忽略标点符号

Question

I have a large dataset that contains keywords, followed eventually by a value. I have managed to read the data in from a pdf format, and am left with data that looks like the following:

  myData <- c("adjuster", "7", "hours", "rate", "oct 2 - 16," , "19", "hours", "rate", "_NA_NA_NA_NA_", "total", "gross", "pay", "6500", "_NA_NA_NA_table",  "NA_copy", "of", "9.16.19 to 9.30.19.xlsx_NA")

myDataDF <- as.data.frame(myData)

My goal is to 'spread' that single column of character data into two columns, one for the alpha values, the second for the numeric values that follow below. I would like to to bring over punctuation, but ignore it as a means of separating keywords from values, as some of the numeric values have punctuation. I would like to collapse (with a space) the keywords, until a numeric value is found, which then is placed in the values column.

I have tried a number of things with this data in different formats (long strings and string splitting), but this format seems the most conducive and clean to get me to the end goal (having data to actually analyze and perform calculations). I just don't know how to qualify keep collapsing until you hit a number in R.

Ultimately, it would be nice if looked as such:

+==========================================+============================+
|                 keyword                  |           value            |
+==========================================+============================+
| adjuster                                 | 7                          |
+------------------------------------------+----------------------------+
| hours rate oct 2 - 16                    | 19                         |
+------------------------------------------+----------------------------+
| hours rate _NA_NA_NA_NA_ total gross pay | 6500                       |
+------------------------------------------+----------------------------+
| _NA_NA_NA_table NA_copy of               | 9.16.19 to 9.30.19.xlsx_NA |
+------------------------------------------+----------------------------+

Answer 1

The last row pattern is not very clear. Based on the data, we could create a grouping column by detecting only numeric values or the 'xlsx' in the 'myData' column and then summarise by paste ing the values except the last and the second column as the last value

library(dplyr)
library(stringr)
myDataDF %>% 
     group_by(grp = lag(cumsum(str_detect(myData, '^\\d+$|xlsx')), 
          default = 0)) %>% 
     summarise(keyword = str_c(myData[-n()], collapse = ' '), 
               value = last(myData), .groups = 'drop') %>% 
     select(-grp)

-output

# A tibble: 4 x 2
#  keyword                                  value                     
#  <chr>                                    <chr>                     
#1 adjuster                                 7                         
#2 hours rate oct 2 - 16,                   19                        
#3 hours rate _NA_NA_NA_NA_ total gross pay 6500                      
#4 _NA_NA_NA_table NA_copy of               9.16.19 to 9.30.19.xlsx_NA