read csv file and filter specific columns in Lambda (Java8)

Currently, after the map operation, you have a Stream<List<String>> and you're trying to compare that with a String, hence will never yield the expected outcome.

Now, to the solution; from what I can gather it seems that you want to retain the entire line if the TIMESTAMP and RESPONSETIME are valid integers.

One way to go about this is:

 List<String> rows = br.lines()
            .skip(1) // skip headers
            .map(s -> new AbstractMap.SimpleEntry<>(s,s.split(DILIMETER)))
            .filter(a -> isInteger(a.getValue()[0]) && isInteger(a.getValue()[2]))
            .map(AbstractMap.SimpleEntry::getKey)
            .collect(Collectors.toList());

and the isInteger function being defined as follows:

public static boolean isInteger(String input)
{
      if(input == null || input.trim().isEmpty()) return false;
      for (char c : input.toCharArray())
          if (!Character.isDigit(c)) return false;
      return true;
}

Another solution being is if you want to retrieve a List<String[]> where each array represents the individual data of each line then you can do:

List<String[]> rows = br.lines()
                .skip(1) // skip headers
                .map(s -> s.split(DILIMETER))
                .filter(a -> isInteger(a[0]) && isInteger(a[2]))
                .collect(Collectors.toList());

Note, if the file being read only contains the data with no headers then there is no need to perform the skip operation.

the problem is you will just get a list full of "TIMESTAMP" which isn't useful. if the file format is always the same, meaning the order and number of headers, you can just skip the first line, then read each of the lines with the data and access just the columns with the data you want to validate. and it may be better to use a for or while loop so you can terminate early.

boolean allNumericData = true;
do{
   String[] row = br.nextLine().split(DELIMITER)
   if(!isNumeric(row[0])||!isNumeric(row[2])){
       allNumericData = false;
   }
}while(allNumericData ||br.nextRow = null)

if the headers can be different then open the file read the first row to determine the index of the data required to validate, and do the same as above but with the found index. also this is pseudo code. you will need to do the validation and handling of fetching and null checking the next row

Presently this is what you doing :

1. read all lines that give : List<String>
2. You split it so that give you String[] and you convert it to List<String> (that is correct) but you are in a map so the result of the map is Stream<List<String>>
3. You filter and a is a List. You try to compare a List and a String. Bip bip bop bop, problem!

Like @YCF_L said, try it without lambda..

You can as well use flatMap and later on filter only string containing digits:

List<String> timeAndResponse = br.lines()
        .flatMap(s -> Arrays.stream(s.split(",")))
        .filter(s -> s.chars().allMatch(Character::isDigit))
        .collect(Collectors.toList());

In this case you are working with streams only:

.flatMap(s -> Arrays.stream(s.split(","))) we take an individual line from the file, split it by , - take a stream from an intermediate array and finally call flatMap. This will give us Stream<String> where String is and individual string from original line 1544190995,US,500,Y. After this point let's just leave only numeric string using filter. Finally let's collect everything to a List, which will contain the following values:

[1544190995, 500, 1723922044, 370, 1711557214, 750]

I hope this helps.