什么是减少数据的重复最简单,最有效的方法。
我试图让一个算法,但它开始得到途径复杂。
我有保持在阵列这样的数据:[[data, 'country_code',value],[data, 'country_code',value],[data, 'country_code',value],[data, 'country_code',value]]
比如我有[[2019-01-23, "GER", 200],[2019-01-23,"USA",300],[2019-01-23,"GER", 301]].
我需要:
[[2019-01-23,"GER", 501],[2019-01-23,"USA",300]]
用defaultdict
积累,并用一个列表理解收集的结果:
>>> from collections import defaultdict
>>> d = defaultdict(int)
>>> for date, code, n in L:
... d[date, code] += n
...
>>> [[date, code, n] for [[date, code], n] in d.items()]
[['2019-01-23', 'GER', 501], ['2019-01-23', 'USA', 300]]
最惯用的方式做到这一点是使用Counter
,从collections
库:
>>> from collections import Counter
>>> data = [
... ['2019-01-23', 'GER', 200],
... ['2019-01-23', 'USA', 300],
... ['2019-01-23', 'GER', 301],
... ]
>>> counter = Counter()
>>> for date, country_code, count in data:
... counter[(date, country_code)] += count
...
>>> counter
Counter({('2019-01-23', 'GER'): 501, ('2019-01-23', 'USA'): 300})
>>> output_data = [[date, country_code, count] for (date, country_code), count in counter.items()]
>>> output_data
[['2019-01-23', 'USA', 300], ['2019-01-23', 'GER', 501]]