Last active
August 29, 2015 14:17
-
-
Save d2fn/71c73960192d03eb3ffe to your computer and use it in GitHub Desktop.
Bucketed top-k in Python using heap-ordered arrays
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| #!/usr/bin/env python | |
| """ | |
| Usage: topk [key-column-name] [score-column-name] | |
| Reads tsv rows, including a header row, from stdin. | |
| Writes to stdout the top 10 rows per distinct value of the given key column. | |
| Author Dietrich Featherston | |
| """ | |
| import sys | |
| import re | |
| import heapq | |
| import string | |
| K = 10 | |
| # push the row into the heap ordered list for the given bucket | |
| # using a heap of bounded size keeps memory use per bucket bound | |
| # and limits time complexity of checking N rows to O(NlogK) | |
| def add(buckets, key, score, row): | |
| h = [] | |
| if key in buckets: | |
| h = buckets[key] | |
| else: | |
| buckets[key] = h | |
| heapq.heappush(h, (score, row)) | |
| if(len(h) > K): | |
| heapq.heappop(h) | |
| def tsv(row): | |
| return string.join(row, '\t') | |
| if __name__ == "__main__": | |
| buckets = {} | |
| columns = [] | |
| first = True | |
| key_column = sys.argv[1] | |
| score_column = sys.argv[2] | |
| key_column_index = 0 | |
| score_column_index = 0 | |
| # read each line | |
| for line in sys.stdin: | |
| row = re.split("\s+", line) | |
| # first row is column names | |
| if first: | |
| columns = row | |
| key_column_index = columns.index(key_column) | |
| score_column_index = columns.index(score_column) | |
| first = False | |
| else: | |
| # compute the row key and score | |
| key = int(row[key_column_index]) | |
| score = int(row[score_column_index]) | |
| # add to heap | |
| add(buckets, key, score, row) | |
| print tsv(columns) | |
| # sort the keys | |
| keys = sorted(buckets, key=lambda key: buckets[key]) | |
| for key in keys: | |
| # get the heap for this key | |
| h = buckets[key] | |
| ordered_list = reversed([heapq.heappop(h) for i in range(len(h))]) | |
| # print each item in the heap in order | |
| for (score, row) in ordered_list: | |
| print tsv(row) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment