CODEWITHSUNDEEP
rstringdata-cleaning
Brad
Brad

Reputation: 680

R: Cleaning a string using a list of wanted substrings

I have a dataframe with a string

Clause <- c('Big Truck Pb Anomaly Low', 'Red Truck Fe Anomaly High', 'Blue Truck Pb Anomaly High & Old Truck Fe Anomaly Low')
Input <- data.frame(Clause)

I would like to clean that string, by retaining only substrings found within a cleaning list;

Keepers <- c('Anomaly', 'Low', 'High', '&', 'Pb', 'Fe', ' ', 'SomethingNotPresent')

The desired result is below.

Wanted <- c('Pb Anomaly Low', 'Fe Anomaly High', 'Pb Anomaly High & Fe Anomaly Low')
Result <- data.frame(Wanted)

Note: The 'Keepers' list will also contain items such as 'SomethingNotPresent'

Upvotes: 1

Views: 23

Answers (2)

Ronak Shah
Ronak Shah

Reputation: 388862

You may split the string at each word and keep only Keepers words for each row.

sapply(strsplit(Input$Clause, '\\s+'), function(x) 
       paste0(x[x %in% Keepers], collapse = ' '))

#[1] "Pb Anomaly Low"     "Fe Anomaly High"     "Anomaly High & Fe Anomaly Low"

Upvotes: 2

Tim Biegeleisen
Tim Biegeleisen

Reputation: 520968

You may form a regex alternation of whitelisted terms to keep. Then use a negative lookahead pattern to identify all terms/whitespace which should be removed:

alternation <- paste(Keepers, collapse="|")
regex <- paste0("\\s*(?!(?:", alternation, "))(?<!\\S)\\S+(?!\\S)\\s*")
df$clause <- gsub("\\s+", " ", trimws(gsub(regex, " ", df$clause, perl=TRUE)))
df

                            clause
1                   Pb Anomaly Low
2                  Fe Anomaly High
3 Pb Anomaly High & Fe Anomaly Low

Data:

inp <- c('Big Truck Pb Anomaly Low', 'Red Truck Fe Anomaly High',
         'Blue Truck Pb Anomaly High & Old Truck Fe Anomaly Low')
df <- data.frame(clause=inp, stringsAsFactors=FALSE)
Keepers <- c('Anomaly', 'Low', 'High', '&', 'Pb', 'Fe', ' ', 'SomethingNotPresent')

Upvotes: 0

Related Questions