Frustation with messy string data - Need to extract all patterns into new columns

[momoandi]

I currently have a column called MDRLOGD_MESSAGE that contains so much information. This one column needs to split in so many ways. Here is a snippet of the data.
df <- data.frame(MDRLOGD_RUN_DATE = Sys.Date() - 1:4,
MDRLOGD_MESSAGE = c('Mapping completed successfully at: 31-MAY-2022 01:44:31 AM','Elapsed time: 2 seconds','Selected: 0 -- Inserted: 0 -- Updated: 0 -- Deleted: 0 -- Errors: 0','Running mapping: DELETE_MST_FACULTY_ATTRIBUTE'))

I would like the separate the numeric information along with its appropriate string information.
df %>%
mutate (Elapsed_Time = ifelse(str_detect(MDRLOGD_MESSAGE, "Elapsed time:"),
str_extract(MDRLOGD_MESSAGE,'(\d\s\b\w+)'), 0), Record_Metric_Name = ifelse(str_detect(MDRLOGD_MESSAGE,"Selected:"), str_extract_all(MDRLOGD_MESSAGE, '(\\b\\w+)'), "NULL"),
Record_Metric_Number = ifelse(str_detect(MDRLOGD_MESSAGE,"Selected:" ),
str_extract_all('\d\s\b\w+&')
Mapping_Run_Date = ifelse(str_detect(MDRLOGD_MESSAGE, "Mapping completed successfully at:"),
str_extract(MDRLOGD_MESSAGE, '?=(:)'),"Null"))

So far, I came up with this to try to manipulate with only successfully splitting up the Elapsed time into its own column. I get an error with the rest: Error in ifelse(., str_detect(MDRLOGD_MESSAGE, "Selected:"), strsplit(MDRLOGD_MESSAGE, :
'list' object cannot be coerced to type 'logical'
Specifically, I would like to separate the 'Selected: 0 -- Inserted: 0 -- Updated: 0 -- Deleted: 0 -- Errors: 0" into their own columns so I can sum up inserted and updated separately by date.

Any help will be appreciated

[technocrat]

See if this helps

# row contains a variable, here called target, that contains
# 4 vectors of strings
(target = c('Mapping completed successfully at: 31-MAY-2022 01:44:31 AM','Elapsed time: 2 seconds','Selected: 0 -- Inserted: 0 -- Updated: 0 -- Deleted: 0 -- Errors: 0','Running mapping: DELETE_MST_FACULTY_ATTRIBUTE'))
#> [1] "Mapping completed successfully at: 31-MAY-2022 01:44:31 AM"         
#> [2] "Elapsed time: 2 seconds"                                            
#> [3] "Selected: 0 -- Inserted: 0 -- Updated: 0 -- Deleted: 0 -- Errors: 0"
#> [4] "Running mapping: DELETE_MST_FACULTY_ATTRIBUTE"

# target[1] contains information success/fail and date time
(outcome <- gsub(" at.*$","",target[1]))
#> [1] "Mapping completed successfully"
(finish  <- gsub(".*at: ","",target[1]) |> lubridate::dmy_hms(x = _))
#> [1] "2022-05-31 01:44:31 UTC"
# target[2] contains a numeric value to be extracted
(elapsed <- gsub("[^.0-9]","",target[2]))
#> [1] "2"
# target[3] contains a classification of mappings
split_on_colon <- function(x) strsplit(x,": ")
(results <- strsplit(target[3]," -- "))
#> [[1]]
#> [1] "Selected: 0" "Inserted: 0" "Updated: 0"  "Deleted: 0"  "Errors: 0"
(results <- lapply(results,split_on_colon) |> unlist(x = _))
#>  [1] "Selected" "0"        "Inserted" "0"        "Updated"  "0"       
#>  [7] "Deleted"  "0"        "Errors"   "0"
(result_types <- results[1:length(results) %% 2 != 0])
#> [1] "Selected" "Inserted" "Updated"  "Deleted"  "Errors"
(result_counts <- as.numeric(results[1:length(results) %% 2 == 0]))
#> [1] 0 0 0 0 0
# target[4] contains the mapping used
(mapping <- gsub("^.*: ","",target[4]))
#> [1] "DELETE_MST_FACULTY_ATTRIBUTE"

^{Created on 2023-01-02 with reprex v2.0.2}

[DavoWW]

Hi @momoandi,
I went a little further than @technocrat and tried to bring all the required output together using dplyr verbs. I also assumed that the run_date was the same for the 4 lines in the dataframe:

suppressPackageStartupMessages(library(tidyverse))

df <- data.frame(MDRLOGD_RUN_DATE = as.Date("2022-06-10"),
                 MDRLOGD_MESSAGE = c('Mapping completed successfully at: 31-MAY-2022 01:44:31 AM',
                                     'Elapsed time: 2 seconds',
                                     'Selected: 0 -- Inserted: 0 -- Updated: 0 -- Deleted: 0 -- Errors: 0',
                                     'Running mapping: DELETE_MST_FACULTY_ATTRIBUTE'))

df %>% 
  mutate(line_index = 1:4) %>% 
  group_by(MDRLOGD_RUN_DATE) %>% 
  pivot_wider(names_from=line_index, values_from = MDRLOGD_MESSAGE) %>% 
  rename(completed = `1`, elapsed = `2`, action = `3`, mapping = `4`) %>% 
  mutate(completed = as.POSIXct(str_remove(completed, "Mapping completed successfully at: "),
                                format="%d-%b-%Y %H:%M:%S %p")) %>% 
  mutate(elapsed = str_remove(elapsed, "Elapsed time: ")) %>%
  mutate(mapping = str_remove(mapping, "Running mapping: ")) %>% 
  separate(action, into=c("selected", "inserted", "updated", "deleted", "errors"), sep="--") %>% 
  map_at(., .at=c("selected", "inserted", "updated", "deleted", "errors"), parse_number) %>% 
  as.data.frame()
#>   MDRLOGD_RUN_DATE           completed   elapsed selected inserted updated
#> 1       2022-06-10 2022-05-31 01:44:31 2 seconds        0        0       0
#>   deleted errors                      mapping
#> 1       0      0 DELETE_MST_FACULTY_ATTRIBUTE

^{Created on 2023-01-03 with reprex v2.0.2}

Hope this helps.

[momoandi]

Hi @DavoWW ,

I appreciate this solution because it uses tidyverse which I am familiar with. I will try this solution to see if this works. Yes, you're assumptions are correct. It is the same date for every row. That was my mistake. Thank you so much.