.csv 處理

這里學(xué)習(xí) 通過一個(gè).csv文件進(jìn)行基本的迭代，來創(chuàng)建字典和收集匯總統(tǒng)計(jì)。
不過總的用一個(gè)詞來描述 這個(gè)csv方法就是 tedious

import csv
%precision 2  #設(shè)置列印的浮點(diǎn)數(shù)據(jù)精度為2。
?
with open('mpg.csv') as csvfile:
    mpg = list(csv.DictReader(csvfile))
  
# 使用csv.DictReader讀取我們的mpg.csv 并將其轉(zhuǎn)換為列表的字典。


mpg[:1] # The first dictionaries in our list.

#輸出如下
>>>
[OrderedDict([('', '1'),
              ('manufacturer', 'audi'),
              ('model', 'a4'),
              ('displ', '1.8'),
              ('year', '1999'),
              ('cyl', '4'),
              ('trans', 'auto(l5)'),
              ('drv', 'f'),
              ('cty', '18'),
              ('hwy', '29'),
              ('fl', 'p'),
              ('class', 'compact')])]

len(mpg)
# 有234 個(gè)字典key mpg[233] 是最后一個(gè)
mpg[233].keys()
>>>
234

keys gives us the column names of our csv.

mpg[233].keys()
odict_keys(['', 'manufacturer', 'model', 'displ', 'year', 'cyl', 'trans', 'drv', 'cty', 'hwy', 'fl', 'class'])

下面這個(gè)是如何找到每個(gè)城市的平均mpg，以及每個(gè)hwy的平均mpg：
因?yàn)樽值淅锏膬?nèi)容都是string，所以需要轉(zhuǎn)化成float才可以計(jì)算

sum(float(d['cty']) for d in mpg) / len(mpg)
sum(float(d['hwy']) for d in mpg) / len(mpg)

現(xiàn)在嘗試返回?cái)?shù)據(jù)組中所有的汽缸的數(shù)據(jù)值:

cylinders = set(d['cyl'] for d in mpg)
cylinders
>>> {'4', '5', '6', '8'}

這里用氣缸的數(shù)量來分組汽車，并找出每個(gè)組的平均mpg。

CtyMpgByCyl = []
?# 創(chuàng)建一個(gè)list

for c in cylinders:                     # 循環(huán)這個(gè)汽缸的list
    summpg = 0
    cyltypecount = 0
    for d in mpg:                       # 迭代所有的字典元素
        if d['cyl'] == c:               # 如果找到了當(dāng)下循環(huán)的汽缸值
            summpg += float(d['cty'])   # 把cty的mpg累加
            cyltypecount += 1           # increment the count
    CtyMpgByCyl.append((c, summpg / cyltypecount)) # append the tuple ('cylinder', 'avg mpg')
?
CtyMpgByCyl.sort(key=lambda x: x[0])  
CtyMpgByCyl
[('4', 21.01), ('5', 20.50), ('6', 16.22), ('8', 12.57)]

其他變量分類的例子：

vehicleclass = set(d['class'] for d in mpg) # what are the class types
vehicleclass
>>> {'2seater', 'compact', 'midsize', 'minivan', 'pickup', 'subcompact', 'suv'}

#average hwy mpg for each class of vehicle


HwyMpgByClass = []

for t in vehicleclass: # iterate over all the vehicle classes
    summpg = 0
    vclasscount = 0
    for d in mpg: # iterate over all dictionaries
        if d['class'] == t: # if the cylinder amount type matches,
            summpg += float(d['hwy']) # add the hwy mpg
            vclasscount += 1 # increment the count
    HwyMpgByClass.append((t, summpg / vclasscount)) # append the tuple ('class', 'avg mpg')

HwyMpgByClass.sort(key=lambda x: x[1])
HwyMpgByClass
HwyMpgByClass = []

for t in vehicleclass: # iterate over all the vehicle classes
    summpg = 0
    vclasscount = 0
    for d in mpg: # iterate over all dictionaries
        if d['class'] == t: # if the cylinder amount type matches,
            summpg += float(d['hwy']) # add the hwy mpg
            vclasscount += 1 # increment the count
    HwyMpgByClass.append((t, summpg / vclasscount)) # append the tuple ('class', 'avg mpg')
?
HwyMpgByClass.sort(key=lambda x: x[1])
HwyMpgByClass
#Output below
[('pickup', 16.88),
 ('suv', 18.13),
 ('minivan', 22.36),
 ('2seater', 24.80),
 ('midsize', 27.29),
 ('subcompact', 28.14),
 ('compact', 28.30)]

time 和 datetime

前提：Python中的一些基本的知識(shí)：
應(yīng)該意識(shí)到該日期和 時(shí)間的存儲(chǔ)有許多不同的方式。
用于存儲(chǔ)日期最常用的傳統(tǒng)方法之一， 時(shí)間在網(wǎng)上系統(tǒng)是基于從紀(jì)元epoch 的偏移量offset。這個(gè)epoch是1970年1月1日。
所以如果看到很大的數(shù)字，而希望看到日期和時(shí)間， 需要轉(zhuǎn)換它們，使數(shù)據(jù)變得有意義。

import datetime as dt
import time as tm

time returns the current time in seconds since the Epoch. (January 1st, 1970)

tm.time()
>>> 1523682711.76

dtnow = dt.datetime.fromtimestamp(tm.time())
#Convert the timestamp to datetime.
dtnow
>>>datetime.datetime(2018, 4, 14, 4, 51, 12, 996246)

#更方便的寫法
dtnow.year, dtnow.month, dtnow.day, dtnow.hour, dtnow.minute, dtnow.second 
# get year, month, day, etc.from a datetime
(2018, 4, 14, 4, 51, 12)

timedelta 是兩個(gè)時(shí)間之間的差值，可以用來計(jì)算前后的時(shí)間，在datetime 包里
datetime 返回的是今天的日期

delta = dt.timedelta(days = 100) # create a timedelta of 100 days
delta
>>> datetime.timedelta(100)


today = dt.date.today()
# 返回一百天前的日期
today - delta   
datetime.date(2018, 1, 4)

#比較日期
today > today-delta    
True