Multi Line Record

September 02, 2021

I was solving a puzzle at https://adventofcode.com/2020/day/4 I needed to read multiple lines and build a record (passport).

After using System.IO.File.ReadLines I found a bug: missing the last record.

Example input:

ecl:gry pid:860033327 eyr:2020 hcl:#fffffd
byr:1937 iyr:2017 cid:147 hgt:183cm

iyr:2013 ecl:amb cid:350 eyr:2023 pid:028048884
hcl:#cfa07d byr:1929

hcl:#ae17e1 iyr:2013
eyr:2024
ecl:brn pid:760753108 byr:1931
hgt:179cm

hcl:#cfa07d eyr:2025 pid:166559648
iyr:2011 ecl:brn hgt:59in

Here’s the initial function that reads the above text and return three passport records.

        private IEnumerable<...> GetPassports()
        {
            var passport = new ...;

            foreach (var line in File.ReadLines(@".\2020\day4\test.txt"))
            {
                if (line.Length == 0) // empty line separates record
                {
                    yield return passport;
                    passoport = new ...;
                    continue;
                }

                // parse line to build passport
            }
        }

The code identifies a passport when a blank line is entered. The last record however is dropped. Fixing this is simple, adding one more yield return

        private IEnumerable<...> GetPassports()
        {
            var passport = new ...;

            foreach (var line in File.ReadLines(@".\2020\day4\test.txt"))
            {
                if (line.Length == 0) // empty line separates record
                {
                    yield return passport;
                    passoport = new ...;
                    continue;
                }

                // parse line to build passport
            }

            yield return passport; // poor fix... I introduced a bug.
        }

However, the new change adds a bug: When no record is read from the source, it returns a passport instead of Enumerable.Empty()

This wasn’t the issue for the puzzle, there’s always a record. However, I wanted to make it robust. Avoid having to remember to apply in two places would also be nice.

I reverted the change.

Then there were two likely option to pursue:

control the input file

Always have a new line in the end. That is, make a blank line mandatory: record lines must include a blank lines.

Instead of changing the file I can do it in the code to ensure the record separator (blank line) to be presented at the end of file read.

        private IEnumerable<...> GetPassports()
        {
            var passport = new ...;

            foreach (var line in
                File.ReadLines(@".\2020\day4\test.txt")
                .Concat(new string[] { string.Empty })) // better, but can be improved
            {
                if (line.Length == 0) // empty line separates record
                {
                    yield return passport;
                    passoport = new ...;
                    continue;
                }

                // parse line to build passport
            }
        }

More edge cases to ponder.

What if the ending is inconsistent? Some input may already have the line ending. Perhaps a better version is to get a set of lines and the extra empty line can be ignored during parsing.
This problem is special in the case of File as EOF terminates the process and exits the loop. Whereas handling a line to parse a record, you have all data from begin and end.