CSV是存儲(chǔ)數(shù)據(jù)的常用方法
Python的卓越靈活性和易用性使其成為最受歡迎的編程語(yǔ)言之一,尤其是對(duì)于數(shù)據(jù)處理和機(jī)器學(xué)習(xí)方面來(lái)說(shuō),其強(qiáng)大的數(shù)據(jù)處理庫(kù)和算法庫(kù)使得python成為入門(mén)數(shù)據(jù)科學(xué)的首選語(yǔ)言。在日常使用中,CSV,JSON和XML三種數(shù)據(jù)格式占據(jù)主導(dǎo)地位。下面我將針對(duì)三種數(shù)據(jù)格式來(lái)分享其快速處理的方法。
CSV數(shù)據(jù)
CSV是存儲(chǔ)數(shù)據(jù)的最常用方法。在Kaggle比賽的大部分?jǐn)?shù)據(jù)都是以這種方式存儲(chǔ)的。我們可以使用內(nèi)置的Python csv庫(kù)來(lái)讀取和寫(xiě)入CSV。通常,我們會(huì)將數(shù)據(jù)讀入列表列表。
看看下面的代碼。當(dāng)我們運(yùn)行csv.reader()所有CSV數(shù)據(jù)變得可訪問(wèn)時(shí)。該csvreader.next()函數(shù)從CSV中讀取一行; 每次調(diào)用它,它都會(huì)移動(dòng)到下一行。我們也可以使用for循環(huán)遍歷csv的每一行for row in csvreader 。確保每行中的列數(shù)相同,否則,在處理列表列表時(shí),最終可能會(huì)遇到一些錯(cuò)誤。
- import csv
- filename = "my_data.csv"
- fields = []
- rows = []
- # Reading csv file
- with open(filename, 'r') as csvfile:
- # Creating a csv reader object
- csvcsvreader = csv.reader(csvfile)
- # Extracting field names in the first row
- fields = csvreader.next()
- # Extracting each data row one by one
- for row in csvreader:
- rows.append(row)
- # Printing out the first 5 rows
- for row in rows[:5]:
- print(row)
在Python中寫(xiě)入CSV同樣容易。在單個(gè)列表中設(shè)置字段名稱,并在列表列表中設(shè)置數(shù)據(jù)。這次我們將創(chuàng)建一個(gè)writer()對(duì)象并使用它將我們的數(shù)據(jù)寫(xiě)入文件,與讀取時(shí)的方法基本一樣。
- import csv
- # Field names
- fields = ['Name', 'Goals', 'Assists', 'Shots']
- # Rows of data in the csv file
- rows = [ ['Emily', '12', '18', '112'],
- ['Katie', '8', '24', '96'],
- ['John', '16', '9', '101'],
- ['Mike', '3', '14', '82']]
- filename = "soccer.csv"
- # Writing to csv file
- with open(filename, 'w+') as csvfile:
- # Creating a csv writer object
- csvcsvwriter = csv.writer(csvfile)
- # Writing the fields
- csvwriter.writerow(fields)
- # Writing the data rows
- csvwriter.writerows(rows)
我們可以使用Pandas將CSV轉(zhuǎn)換為快速單行的字典列表。將數(shù)據(jù)格式化為字典列表后,我們將使用該dicttoxml庫(kù)將其轉(zhuǎn)換為XML格式。我們還將其保存為JSON文件!
- import pandas as pd
- from dicttoxml import dicttoxml
- import json
- # Building our dataframe
- data = {'Name': ['Emily', 'Katie', 'John', 'Mike'],
- 'Goals': [12, 8, 16, 3],
- 'Assists': [18, 24, 9, 14],
- 'Shots': [112, 96, 101, 82]
- }
- df = pd.DataFrame(data, columns=data.keys())
- # Converting the dataframe to a dictionary
- # Then save it to file
- data_dict = df.to_dict(orient="records")
- with open('output.json', "w+") as f:
- json.dump(data_dict, f, indent=4)
- # Converting the dataframe to XML
- # Then save it to file
- xml_data = dicttoxml(data_dict).decode()
- with open("output.xml", "w+") as f:
- f.write(xml_data)
JSON數(shù)據(jù)
JSON提供了一種簡(jiǎn)潔且易于閱讀的格式,它保持了字典式結(jié)構(gòu)。就像CSV一樣,Python有一個(gè)內(nèi)置的JSON模塊,使閱讀和寫(xiě)作變得非常簡(jiǎn)單!我們以字典的形式讀取CSV時(shí),然后我們將該字典格式數(shù)據(jù)寫(xiě)入文件。
- import json
- import pandas as pd
- # Read the data from file
- # We now have a Python dictionary
- with open('data.json') as f:
- data_listofdict = json.load(f)
- # We can do the same thing with pandas
- data_df = pd.read_json('data.json', orient='records')
- # We can write a dictionary to JSON like so
- # Use 'indent' and 'sort_keys' to make the JSON
- # file look nice
- with open('new_data.json', 'w+') as json_file:
- json.dump(data_listofdict, json_file, indent=4, sort_keys=True)
- # And again the same thing with pandas
- export = data_df.to_json('new_data.json', orient='records')
正如我們之前看到的,一旦我們獲得了數(shù)據(jù),就可以通過(guò)pandas或使用內(nèi)置的Python CSV模塊輕松轉(zhuǎn)換為CSV。轉(zhuǎn)換為XML時(shí),可以使用dicttoxml庫(kù)。具體代碼如下:
- import json
- import pandas as pd
- import csv
- # Read the data from file
- # We now have a Python dictionary
- with open('data.json') as f:
- data_listofdict = json.load(f)
- # Writing a list of dicts to CSV
- keys = data_listofdict[0].keys()
- with open('saved_data.csv', 'wb') as output_file:
- dict_writer = csv.DictWriter(output_file, keys)
- dict_writer.writeheader()
- dict_writer.writerows(data_listofdict)
XML數(shù)據(jù)
XML與CSV和JSON有點(diǎn)不同。CSV和JSON由于其既簡(jiǎn)單又快速,可以方便人們進(jìn)行閱讀,編寫(xiě)和解釋。而XML占用更多的內(nèi)存空間,傳送和儲(chǔ)存需要更大的帶寬,更多存儲(chǔ)空間和更久的運(yùn)行時(shí)間。但是XML也有一些基于JSON和CSV的額外功能:您可以使用命名空間來(lái)構(gòu)建和共享結(jié)構(gòu)標(biāo)準(zhǔn),更好地傳承,以及使用XML、DTD等數(shù)據(jù)表示的行業(yè)標(biāo)準(zhǔn)化方法。
要讀入XML數(shù)據(jù),我們將使用Python的內(nèi)置XML模塊和子模ElementTree。我們可以使用xmltodict庫(kù)將ElementTree對(duì)象轉(zhuǎn)換為字典。一旦我們有了字典,我們就可以轉(zhuǎn)換為CSV,JSON或Pandas Dataframe!具體代碼如下:
- import xml.etree.ElementTree as ET
- import xmltodict
- import json
- tree = ET.parse('output.xml')
- xml_data = tree.getroot()
- xmlstr = ET.tostring(xml_data, encoding='utf8', method='xml')
- data_dict = dict(xmltodict.parse(xmlstr))
- print(data_dict)
- with open('new_data_2.json', 'w+') as json_file:
- json.dump(data_dict, json_file, indent=4, sort_keys=True)