您好,登錄后才能下訂單哦!
本篇內(nèi)容主要講解“Python如何實(shí)現(xiàn)SQL自動(dòng)化”,感興趣的朋友不妨來看看。本文介紹的方法操作簡單快捷,實(shí)用性強(qiáng)。下面就讓小編來帶大家學(xué)習(xí)“Python如何實(shí)現(xiàn)SQL自動(dòng)化”吧!
從基礎(chǔ)開始
import pyodbc from datetime import datetime classSql: def__init__(self, database, server="XXVIR00012,55000"): # here we are telling python what to connect to (our SQL Server) self.cnxn = pyodbc.connect("Driver={SQL Server Native Client 11.0};" "Server="+server+";" "Database="+database+";" "Trusted_Connection=yes;") # initialise query attribute self.query ="-- {}\n\n-- Made in Python".format(datetime.now() .strftime("%d/%m/%Y"))
這個(gè)代碼就是操作MS SQL服務(wù)器的基礎(chǔ)。只要編寫好這個(gè)代碼,通過Python 連接到SQL 僅需:
sql = Sql('database123')
很簡單對(duì)么?同時(shí)發(fā)生了幾件事,下面將對(duì)此代碼進(jìn)行剖析。class Sql:
首先要注意,這個(gè)代碼包含在一個(gè)類中。筆者發(fā)現(xiàn)這是合乎邏輯的,因?yàn)樵诖烁袷街?,已?jīng)對(duì)此特定數(shù)據(jù)庫進(jìn)行了增添或移除進(jìn)程。若見其工作過程,思路便能更加清晰。
初始化類:
def __init__(self, database,server="XXVIR00012,55000"):
因?yàn)楣P者和同事幾乎總是連接到相同的服務(wù)器,所以筆者將這個(gè)通用瀏覽器的名稱設(shè)為默認(rèn)參數(shù)server。
在“Connect to Server”對(duì)話框或者M(jìn)S SQL Server Management Studio的視窗頂端可以找到服務(wù)器的名稱:
下一步,連接SQL:
self.cnxn =pyodbc.connect("Driver={SQL Server Native Client 11.0};" "Server="+self.server+";" "Database="+self.database+";" "Trusted_Connection=yes;")
pyodbc 模塊,使得這一步驟異常簡單。只需將連接字符串過渡到 pyodbc.connect(...) 函數(shù)即可,點(diǎn)擊以了解詳情here。
最后,筆者通常會(huì)在 Sql 類中編寫一個(gè)查詢字符串,sql類會(huì)隨每個(gè)傳遞給類的查詢而更新:
self.query = "-- {}\n\n--Made in Python".format(datetime.now() .strftime("%d/%m/%Y"))
這樣便于記錄代碼,同時(shí)也使輸出更為可讀,讓他人讀起來更舒服。
請(qǐng)注意在下列的代碼片段中,筆者將不再更新代碼中的self.query 部分。
組塊
一些重要函數(shù)非常有用,筆者幾乎每天都會(huì)使用。這些函數(shù)都側(cè)重于將數(shù)據(jù)從數(shù)據(jù)庫中傳入或傳出。
以下圖文件目錄為始:
對(duì)于當(dāng)前此項(xiàng)目,需要:
將文件導(dǎo)入SQL
將其合并到單一表格內(nèi)
根據(jù)列中類別靈活創(chuàng)建多個(gè)表格
SQL類不斷被充實(shí)后,后續(xù)會(huì)容易很多:
import sys sys.path.insert(0, r'C:\\User\medium\pysqlplus\lib') import os from data importSql sql =Sql('database123') # initialise the Sql object directory =r'C:\\User\medium\data\\' # this is where our generic data is stored file_list = os.listdir(directory) # get a list of all files for file in file_list: # loop to import files to sql df = pd.read_csv(directory+file) # read file to dataframe sql.push_dataframe(df, file[:-4]) # now we convert our file_list names into the table names we have imported to SQL table_names = [x[:-4] for x in file_list] sql.union(table_names, 'generic_jan') # union our files into one new table called 'generic_jan' sql.drop(table_names) # drop our original tables as we now have full table # get list of categories in colX, eg ['hr', 'finance', 'tech', 'c_suite'] sets =list(sql.manual("SELECT colX AS 'category' FROM generic_jan GROUP BY colX", response=True)['category']) for category in sets: sql.manual("SELECT * INTO generic_jan_"+category+" FROM generic_jan WHERE colX = '"+category+"'")
從頭開始。
入棧數(shù)據(jù)結(jié)構(gòu)
defpush_dataframe(self, data, table="raw_data", batchsize=500): # create execution cursor cursor = self.cnxn.cursor() # activate fast execute cursor.fast_executemany =True # create create table statement query ="CREATE TABLE ["+ table +"] (\n" # iterate through each column to be included in create table statement for i inrange(len(list(data))): query +="\t[{}] varchar(255)".format(list(data)[i]) # add column (everything is varchar for now) # append correct connection/end statement code if i !=len(list(data))-1: query +=",\n" else: query +="\n);" cursor.execute(query) # execute the create table statement self.cnxn.commit() # commit changes # append query to our SQL code logger self.query += ("\n\n-- create table\n"+ query) # insert the data in batches query = ("INSERT INTO [{}] ({})\n".format(table, '['+'], [' # get columns .join(list(data)) +']') + "VALUES\n(?{})".format(", ?"*(len(list(data))-1))) # insert data into target table in batches of 'batchsize' for i inrange(0, len(data), batchsize): if i+batchsize >len(data): batch = data[i: len(data)].values.tolist() else: batch = data[i: i+batchsize].values.tolist() # execute batch insert cursor.executemany(query, batch) # commit insert to SQL Server self.cnxn.commit()
此函數(shù)包含在SQL類中,能輕松將Pandas dataframe插入SQL數(shù)據(jù)庫。
其在需要上傳大量文件時(shí)非常有用。然而,Python能將數(shù)據(jù)插入到SQL的真正原因在于其靈活性。
要橫跨一打Excel工作簿才能在SQL中插入特定標(biāo)簽真的很糟心。但有Python在,小菜一碟。如今已經(jīng)構(gòu)建起了一個(gè)可以使用Python讀取標(biāo)簽的函數(shù),還能將標(biāo)簽插入到SQL中。
Manual(函數(shù))
defmanual(self, query, response=False): cursor = self.cnxn.cursor() # create execution cursor if response: returnread_sql(query, self.cnxn) # get sql query output to dataframe try: cursor.execute(query) # execute except pyodbc.ProgrammingErroras error: print("Warning:\n{}".format(error)) # print error as a warning self.cnxn.commit() # commit query to SQL Server return"Query complete."
此函數(shù)實(shí)際上應(yīng)用在union 和 drop 函數(shù)中。僅能使處理SQL代碼變得盡可能簡單。
response參數(shù)能將查詢輸出解壓到DataFrame。generic_jan 表中的colX ,可供摘錄所有獨(dú)特值,操作如下:
sets =list(sql.manual("SELECT colX AS 'category' FROM generic_jan GROUP BYcolX", response=True)['category'])
Union(函數(shù))
構(gòu)建 了manual 函數(shù),創(chuàng)建 union 函數(shù)就簡單了:
defunion(self, table_list, name="union", join="UNION"): # initialise the query query ="SELECT * INTO ["+name+"] FROM (\n" # build the SQL query query +=f'\n{join}\n'.join( [f'SELECT [{x}].* FROM [{x}]'for x in table_list] ) query +=") x" # add end of query self.manual(query, fast=True) # fast execute
創(chuàng)建 union 函數(shù)只不過是在循環(huán)參考 table_list提出的表名,從而為給定的表名構(gòu)建 UNION函數(shù)查詢。然后用self.manual(query)處理。
Drop(函數(shù))
上傳大量表到SQL服務(wù)器是可行的。雖然可行,但會(huì)使數(shù)據(jù)庫迅速過載。 為解決這一問題,需要?jiǎng)?chuàng)建一個(gè)drop函數(shù):
defdrop(self, tables): # check if single or list ifisinstance(tables, str): # if single string, convert to single item in list for for-loop tables = [tables] for table in tables: # check for pre-existing table and delete if present query = ("IF OBJECT_ID ('["+table+"]', 'U') IS NOT NULL " "DROP TABLE ["+table+"]") self.manual(query) # execute
到此,相信大家對(duì)“Python如何實(shí)現(xiàn)SQL自動(dòng)化”有了更深的了解,不妨來實(shí)際操作一番吧!這里是億速云網(wǎng)站,更多相關(guān)內(nèi)容可以進(jìn)入相關(guān)頻道進(jìn)行查詢,關(guān)注我們,繼續(xù)學(xué)習(xí)!
免責(zé)聲明:本站發(fā)布的內(nèi)容(圖片、視頻和文字)以原創(chuàng)、轉(zhuǎn)載和分享為主,文章觀點(diǎn)不代表本網(wǎng)站立場,如果涉及侵權(quán)請(qǐng)聯(lián)系站長郵箱:is@yisu.com進(jìn)行舉報(bào),并提供相關(guān)證據(jù),一經(jīng)查實(shí),將立刻刪除涉嫌侵權(quán)內(nèi)容。