bovespa.py

import datetime

# figure out how many different companies were traded on the Brazilian stock exchange per month

# data: http://www.bmfbovespa.com.br/en_us/services/market-data/historical-data/equities/historical-data/
# file format: http://www.bmfbovespa.com.br/en_us/services/market-data/historical-data/equities/historical-data-1/

class QuoteCache:
    def __init__(self):
        self.stocks = {}

    def read_quote_file(self, year):
        base_filename = 'COTAHIST_A%04d.TXT' % year
        with open(base_filename, 'r') as f:
            for line in f:
                self._parse_quote_line(line)

    def _to_datetime(self, s):
        y, m, d = int(s[:4]), int(s[4:6]), int(s[6:])
        return datetime.date(y, m, d)

    def _parse_quote_line(self, line):
        if line[:2] != '01':
            return

        # ignore fundos de investimento ou BDRs
        spec = line[39:48].strip()
        spec_parts = spec.split()
        if len(spec_parts) > 0:
            share_type = spec_parts[0]
            if share_type in ('CI', 'DRN'):
                return
        
        # so' mercado 'a vista
        market_type = line[24:27]
        if market_type != '010':
            return

        ticker = line[12:24].strip()
        date = self._to_datetime(line[2:10])
        year_month = (date.year, date.month)
        self.stocks.setdefault(year_month, set()).add(ticker[:4])

cache = QuoteCache()
for year in range(1986, 2019):
    cache.read_quote_file(year)

for year_month in sorted(cache.stocks.keys()):
    print "%04d-%02d\t%d" % (year_month[0], year_month[1], len(cache.stocks[year_month]))