昨晚我问了一个类似的问题，但是我的教授对她如何回答这个问题做了一些澄清，这让我感到厌恶。
我有一个3列的csv文件。我将它们保存为字典，但我正试图找到一种方法来阅读year和title_field并找到一个特定的title_field（占用住房单位），与最早的一年（2008年）匹配并取数字在那些旁边的value专栏中，并将其与相同的year（占用住房单位）匹配到NEXT title_field（2009），找到这两个值之间的差异，并打印结果并对2009年和2010年等进行相同操作。像这样：
2008-2009 795
2009-2010 5091
etc.
csv看起来像这样：
year,title_field,value
2014,Total Housing Units,49109
2014,Vacant Housing Units,2814
2014,Occupied Housing Units,46295
2013,Total Housing Units,47888
2013,Vacant Housing Units,4215
2013,Occupied Housing Units,43673
2012,Total Housing Units,45121
2012,Vacant Housing Units,3013
2012,Occupied Housing Units,42108
2011,Total Housing Units,44917
2011,Vacant Housing Units,4213
2011,Occupied Housing Units,40704
2010,Total Housing Units,44642
2010,Vacant Housing Units,3635
2010,Occupied Housing Units,41007
2009,Total Housing Units,39499
2009,Vacant Housing Units,3583
2009,Occupied Housing Units,35916
2008,Total Housing Units,41194
2008,Vacant Housing Units,4483
2008,Occupied Housing Units,36711
我到目前为止的代码是：
import csv
def process(year, field_name, value):
print(year, field_name, value)
with open('denton_housing.csv', 'r', encoding='utf8',newline='') as f:
reader = csv.DictReader(f, delimiter=',')
housing_stats = []
for row in reader:
year = row["year"]
field_name = row["title_field"]
value = int(row["value"])
denton_dict = {"year": year, "field_name": field_name, "value": value}
housing_stats.append(denton_dict)
process(year, field_name, value)
谢谢！我是编程的新手，而且我是一个年长的老兄。我喜欢编程社区无济于事，好像你们都欢迎所有人进入邪教组织（好的一类？）。
你可以这样做：
lists的dicts，其中包含目标title_field值。itertools配方为pairwise()生成器处理排序的list中的每对行/年。实施上述代码：
import csv
from itertools import tee
# From https://docs.python.org/3/library/itertools.html#recipes
def pairwise(iterable):
"s -> (s0,s1), (s1,s2), (s2, s3), ..."
a, b = tee(iterable)
next(b, None)
return zip(a, b)
target_title_field = 'Occupied Housing Units'
csv_filename = 'denton_housing.csv'
with open(csv_filename, 'r', encoding='utf8', newline='') as f:
housing_stats = []
for row in csv.DictReader(f, delimiter=','):
if row['title_field'] == target_title_field:
year = int(row["year"])
field_name = row["title_field"]
value = int(row["value"])
denton_dict = {"year": year, "field_name": field_name, "value": value}
housing_stats.append(denton_dict)
housing_stats.sort(key=lambda row: row['year'])
for r1, r2 in pairwise(housing_stats):
print('{}-{} {:5}'.format(r1['year'], r2['year'], abs(r2['value'] - r1['value'])))
输出：
2008-2009 795
2009-2010 5091
2010-2011 303
2011-2012 1404
2012-2013 1565
2013-2014 2622
一种简单的方法是使用3个列表（每个title_field）来保存年份和值字段，然后您可以处理每个列表。
total = []
vacant = []
occupied = []
with open('denton_housing.csv', 'r', encoding='utf8',newline='') as f:
spamreader = csv.reader(f, delimiter=',')
for row in spamreader:
if row[1] == 'Occupied Housing Units':
# use the data structure you preferred, in this example I use tuple
mytuple = (row[0], row[2])
occupied.append(mytuple)
# do the same for total and vacant list, ignore if you don't need
...
# then you can process the list, for example, occupied
# I assume your csv file is sorted by year, so you may safely assume that each
# year field of the data entry in the occupied list is sorted as well
for i in range(len(occupied)-1):
# if your data contains every year, ie 2008-2014 without missing any
# the year field is useless in this case, so you can just
value_diff = abs(occupied[i][1] - occupied[i+1][1])
# if the year entry is not sorted, and it may missed some years
occupied.sort(key=lambda x: x[0]) # this sort in ascending order
for i in range(len(occupied)-1):
this_year = occupied[i][0]
next_year = occupied[i+1][0]
if next_year - this_year == 1:
value_diff = abs(occupied[i][1] - occupied[i+1][1])
我建议你用熊猫做这件事。然后你可以轻而易举地使用groupby和聚合。
像这样：
df.groupby(df['year'].dt.year)['a'].agg(['value'])
结果：
2012 14
2015 6