Spark filtering with regex

Question

I am trying to filter file data into good and bad data per the date, hence will get 2 result files. From test file, first 4 lines need to go in good data and last 2 lines in bad data.

I am having 2 issues

1. I am not getting any good data, result file is empty

2. and bad data result looks like following - picking up the name characters only

   (,C,h) (,J,u) (,T,h) (,J,o) (,N,e) (,B,i)

Test file

Christopher|Jan 11, 2017|5 
Justin|11 Jan, 2017|5 
Thomas|6/17/2017|5 
John|11-08-2017|5 
Neli|2016|5 
Bilu||5

Load and RDD

scala> val file = sc.textFile("test/data.txt")
scala> val fileRDD = file.map(x => x.split("|"))

RegEx

scala> val singleReg = """(\w(3))\s(\d+)(,)\s(\d(4))|(\d+)\s(\w(3))(,)\s(\d(4))|(\d+)(\/)(\d+)(\/)(\d(4))|(\d+)(-)(\d+)(-)(\d(4))""".r

Is three " (double quotes) in the beginning and end and .r important here?

Filter issue area

scala> val validSingleRecords = fileRDD.filter(x => (singleReg.pattern.matcher(x(1)).matches))
scala> val badSingleRecords = fileRDD.filter(x => !(singleReg.pattern.matcher(x(1)).matches))

Turn array into string

scala> val validSingle = validSingleRecords.map(x => (x(0),x(1),x(2)))
scala> val badSingle = badSingleRecords.map(x => (x(0),x(1),x(2)))

Write file

scala> validSingle.repartition(1).saveAsTextFile("data/singValid")
scala> badSingle.repartition(1).saveAsTextFile("data/singBad")

Update 1 My regex above was wrong, i have updated it as. in scala backslash is a escape character, so need to duplicate

val singleReg = """\\w{3}\\s\\d+,\\s\\d{4}|\\d+\\s\\w{3},\\s\\d{4}|\\d+\/\\d+\/\\d{4}|\\d+-\\d+-\\d{4}""".r

Checked the regex on regex101 and the dates in the first 4 lines pass.

I have run the the test again and i am still getting the same result.

Answer 1

There are 2 issues with the code:

1. The character that you are using to split the lines of data.txt is wrong. It should be '|' instead of "|".
2. The regex singleReg is wrong.

The correct code is as follows:

Load and RDD

scala> val file = sc.textFile("test/data.txt")
scala> val fileRDD = file.map(x => x.split('|'))

RegEx

scala> val singleReg = """\w{3}\s\d{2},\s\d{4}|\d{2}\s\w{3},\s\d{4}|\d{1}\/\d{2}\/\d{4}|\d{2}-\d{2}-\d{4}""".r

Filter

scala> val validSingleRecords = fileRDD.filter(x => (singleReg.pattern.matcher(x(1)).matches))
scala> val badSingleRecords = fileRDD.filter(x => !(singleReg.pattern.matcher(x(1)).matches))

Turn array into string

scala> val validSingle = validSingleRecords.map(x => (x(0),x(1),x(2)))
scala> val badSingle = badSingleRecords.map(x => (x(0),x(1),x(2)))

Write file

scala> validSingle.repartition(1).saveAsTextFile("data/singValid")
scala> badSingle.repartition(1).saveAsTextFile("data/singBad")

The above code will give you following output -

data/singValid

(Christopher,Jan 11, 2017,5 )
(Justin,11 Jan, 2017,5 )
(Thomas,6/17/2017,5 )
(John,11-08-2017,5 )

data/singBad

(Neli,2016,5 )
(Bilu,,5)