Finish Property Sales parser

[AKST]

Some of the data between June & October in 2001 is inconsistently formatted within the same file. So

• There'll be files where D or C rows have X & Y number of columns
• But within the same file they will have a different number of columns for the same row type.

Atm it's currently being ignored, and it may be more than just D and C rows, I don't really recall.

https://github.com/AKST/Australian-Address-Boundaries-Land-Property-Price-Database/blob/0d7b83cd848593c378688c105ca96eac8eb14c04/lib/nsw_vg/property_sale_parser.py#L84-L85

Soultion

The least disruptive method of updating this is better identifying the start and end of rows. Something:

• Read until you see \n{A,B,C,D,Z}; or something as a regex
• Then split by semicolon
• Confirm it has the correct number of rows
  • This is worth doing to ensure we are making correct assumptions about the data we are reading.

I tried something like this earlier but it didn't work.