Here we go then, the 1st day of 2023’s Advent of Code.

Day 1 looked deceptively simple, with part 1 completed in pretty quick time. The only real problem was the large amount of data caused memory overflow issues, so I split it into 4 separate parts for the data statements.

I’ll list some code segments with explanation on how I did it below, but I don’t want to post the answers for this current year on here for obvious reasons.

First off, declaring the variables:

I used two subroutines to do the number crunching. It’s not very complicated, this one is used to find the left most number. It’s first set to 100 so we have a benchmark to compare against later on. It then simply loops through each character based on the length of the string and compares it to a digit.

LN (Left Number) is set to a multiple of 10 to satisfy the problem – so if 3 is found first, it needs to be set to 30 as we want to add the single last digit of the string to it. i.e. if the last digit found is 5 then the value for this line would be 35.

Once a left number is found, the loop obviously needs to exit to prevent it finding any more numbers – so we exit as soon as LN is less than 100 (i.e. any of the other values).

```
SUB FINDNUMSL (TEST$ AS STRING * 96) STATIC
L=LEN(TEST$)
LN = 100 'set to an unfindable value - only required for LN
FOR I=0 TO L-1
M=MID$(TEST$, I, 1)
' Test for left numeric only
IF M="0" THEN
LN=0
END IF
IF M="1" THEN
LN=10
END IF
IF M="2" THEN
LN=20
END IF
IF M="3" THEN
LN=30
END IF
IF M="4" THEN
LN=40
END IF
IF M="5" THEN
LN=50
END IF
IF M="6" THEN
LN=60
END IF
IF M="7" THEN
LN=70
END IF
IF M="8" THEN
LN=80
END IF
IF M="9" THEN
LN=90
END IF
IF LN < 100 THEN ' We need to exit the loop as soon as a numeric is found
EXIT FOR
END IF
NEXT I
' Left most digit found
' End subroutine
END SUB
```

The subroutine for the right most number, imaginatively called FINDNUMSR, is almost identical except for 2 differences. We set RN to the single digit as opposed to a multiple of 10, as that will be added to LN, and we *don’t* exit the loop, meaning we keep going until the end so that RN holds the last value in that string.

With the SUBS setup it’s just a matter of cycling through each input line, calling them both, and totalling up the values as we go.

And as I said, I did this 3 more times for the other 750 lines of data, which yes is a bit cheaty but then XC=BASIC makes me lazy.

Part 2 was trickier as we had to find actual strings of digits that would take precedent over the digits themselves, so if the string looked something like ‘iuskj**nine**ij5hkjlcnd6slkj8klmss**two**‘ the result we want is 92.

To accomplish this I just did more of the same and added extra logic into both subroutines, again exiting the loop early for the left number and letting it complete for the right number.

```
'Part 2
IF M = "o" AND MID$(TEST$,I+1,1) = "n" AND MID$(TEST$,I+2,1) = "e" THEN
LN=10
END IF
IF M = "t" AND MID$(TEST$,I+1,1) = "w" AND MID$(TEST$,I+2,1) = "o" THEN
LN=20
END IF
IF M = "t" AND MID$(TEST$,I+1,1) = "h" AND MID$(TEST$,I+2,1) = "r" AND MID$(TEST$,I+3,1) = "e" AND MID$(TEST$,I+4,1) = "e" THEN
LN=30
END IF
IF M = "f" AND MID$(TEST$,I+1,1) = "o" AND MID$(TEST$,I+2,1) = "u" AND MID$(TEST$,I+3,1) = "r" THEN
LN=40
END IF
IF M = "f" AND MID$(TEST$,I+1,1) = "i" AND MID$(TEST$,I+2,1) = "v" AND MID$(TEST$,I+3,1) = "e" THEN
LN=50
END IF
IF M = "s" AND MID$(TEST$,I+1,1) = "i" AND MID$(TEST$,I+2,1) = "x" THEN
LN=60
END IF
IF M = "s" AND MID$(TEST$,I+1,1) = "e" AND MID$(TEST$,I+2,1) = "v" AND MID$(TEST$,I+3,1) = "e" AND MID$(TEST$,I+4,1) = "n" THEN
LN=70
END IF
IF M = "e" AND MID$(TEST$,I+1,1) = "i" AND MID$(TEST$,I+2,1) = "g" AND MID$(TEST$,I+3,1) = "h" AND MID$(TEST$,I+4,1) = "t" THEN
LN=80
END IF
IF M = "n" AND MID$(TEST$,I+1,1) = "i" AND MID$(TEST$,I+2,1) = "n" AND MID$(TEST$,I+3,1) = "e" THEN
LN=90
END IF
' End of part 2
```

The total run time, for both part 1 and part 2 in the same script was 12 minutes 56 seconds.

This screen shows just 1 part completing – there were 3 more after this and I totalled up the output for my answers.

All files, .bas and .prg, can be found here: https://github.com/jonnypencils/aoc2023