pandas 甯哥敤鍔熻兘

Posted 2020-11-30

鏍囩锛?a href='http://www.mamicode.com/so/1/%e7%9b%b8%e5%85%b3%e7%b3%bb%e6%95%b0' title='鐩稿叧绯绘暟'>鐩稿叧绯绘暟   taf   pen   json鏍煎紡   href   榛樿   娓呯悊   

瀵煎叆鏁版嵁

• pd.read_csv(filename)锛氫粠CSV鏂囦欢瀵煎叆鏁版嵁
• pd.read_table(filename)锛氫粠闄愬畾鍒嗛殧绗︾殑鏂囨湰鏂囦欢瀵煎叆鏁版嵁
• pd.read_excel(filename)锛氫粠Excel鏂囦欢瀵煎叆鏁版嵁
• pd.read_sql(query, connection_object)锛氫粠SQL琛?搴撳鍏ユ暟鎹?/li>
• pd.read_json(json_string)锛氫粠JSON鏍煎紡鐨勫瓧绗︿覆瀵煎叆鏁版嵁
• pd.read_html(url)锛氳В鏋怳RL銆佸瓧绗︿覆鎴栬€匟TML鏂囦欢锛屾娊鍙栧叾涓殑tables琛ㄦ牸
• pd.read_clipboard()锛氫粠浣犵殑绮樿创鏉胯幏鍙栧唴瀹癸紝骞朵紶缁檙ead_table()
• pd.DataFrame(dict)锛氫粠瀛楀吀瀵硅薄瀵煎叆鏁版嵁锛孠ey鏄垪鍚嶏紝Value鏄暟鎹?/li>

瀵煎嚭鏁版嵁

• df.to_csv(filename)锛氬鍑烘暟鎹埌CSV鏂囦欢
• df.to_excel(filename)锛氬鍑烘暟鎹埌Excel鏂囦欢
• df.to_sql(table_name, connection_object)锛氬鍑烘暟鎹埌SQL琛?/li>
• df.to_json(filename)锛氫互Json鏍煎紡瀵煎嚭鏁版嵁鍒版枃鏈枃浠?/li>

鍒涘缓娴嬭瘯瀵硅薄

• pd.DataFrame(np.random.rand(20,5))锛氬垱寤?0琛?鍒楃殑闅忔満鏁扮粍鎴愮殑DataFrame瀵硅薄
• pd.Series(my_list)锛氫粠鍙凯浠ｅ璞y_list鍒涘缓涓€涓猄eries瀵硅薄
• df.index = pd.date_range(鈥?900/1/30鈥? periods=df.shape[0])锛氬鍔犱竴涓棩鏈熺储寮?/li>

鏌ョ湅銆佹鏌ユ暟鎹?/h2>

• df.head(n)锛氭煡鐪婦ataFrame瀵硅薄鐨勫墠n琛?/li>
• df.tail(n)锛氭煡鐪婦ataFrame瀵硅薄鐨勬渶鍚巒琛?/li>
• df.shape()锛氭煡鐪嬭鏁板拰鍒楁暟
• [http://df.info()](https://link.zhihu.com/?target=http%3A//df.info())锛氭煡鐪嬬储寮曘€佹暟鎹被鍨嬪拰鍐呭瓨淇℃伅
• df.describe()锛氭煡鐪嬫暟鍊煎瀷鍒楃殑姹囨€荤粺璁?/li>
• s.value_counts(dropna=False)锛氭煡鐪婼eries瀵硅薄鐨勫敮涓€鍊煎拰璁℃暟
• df.apply(pd.Series.value_counts)锛氭煡鐪婦ataFrame瀵硅薄涓瘡涓€鍒楃殑鍞竴鍊煎拰璁℃暟

鏁版嵁閫夊彇

• df[col]锛氭牴鎹垪鍚嶏紝骞朵互Series鐨勫舰寮忚繑鍥炲垪
• df[[col1, col2]]锛氫互DataFrame褰㈠紡杩斿洖澶氬垪
• s.iloc[0]锛氭寜浣嶇疆閫夊彇鏁版嵁
• s.loc[鈥榠ndex_one鈥榏锛氭寜绱㈠紩閫夊彇鏁版嵁
• df.iloc[0,:]锛氳繑鍥炵涓€琛?/li>
• df.iloc[0,0]锛氳繑鍥炵涓€鍒楃殑绗竴涓厓绱?/li>

鏁版嵁娓呯悊

• df.coumns = [鈥榓鈥?鈥榖鈥?鈥榗鈥榏锛氶噸鍛藉悕鍒楀悕
• pd.isnull()锛氭鏌ataFrame瀵硅薄涓殑绌哄€硷紝骞惰繑鍥炰竴涓狟oolean鏁扮粍
• pd.notnull()锛氭鏌ataFrame瀵硅薄涓殑闈炵┖鍊硷紝骞惰繑鍥炰竴涓狟oolean鏁扮粍
• df.dropna()锛氬垹闄ゆ墍鏈夊寘鍚┖鍊肩殑琛?/li>
• df.dropna(axis=1)锛氬垹闄ゆ墍鏈夊寘鍚┖鍊肩殑鍒?/li>
• df.dropna(axis=1,thresh=n)锛氬垹闄ゆ墍鏈夊皬浜巒涓潪绌哄€肩殑琛?/li>
• df.fillna(x)锛氱敤x鏇挎崲DataFrame瀵硅薄涓墍鏈夌殑绌哄€?/li>
• s.astype(float)锛氬皢Series涓殑鏁版嵁绫诲瀷鏇存敼涓篺loat绫诲瀷
• s.replace(1,鈥榦ne鈥?锛氱敤鈥榦ne鈥欎唬鏇挎墍鏈夌瓑浜?鐨勫€?/li>
• s.replace([1,3],[鈥榦ne鈥?鈥榯hree鈥榏)锛氱敤鈥榦ne鈥樹唬鏇?锛岀敤鈥榯hree鈥樹唬鏇?
• df.rename(columns=lambda x: x + 1)锛氭壒閲忔洿鏀瑰垪鍚?/li>
• df.rename(columns={鈥榦ldname鈥? 鈥榥ew name鈥榼)锛氶€夋嫨鎬ф洿鏀瑰垪鍚?/li>
• df.set_index(鈥榗olumn_one鈥?锛氭洿鏀圭储寮曞垪
• df.rename(index=lambda x: x + 1)锛氭壒閲忛噸鍛藉悕绱㈠紩

鏁版嵁澶勭悊锛欶ilter銆丼ort鍜孏roupBy

• df[df[col] > 0.5]锛氶€夋嫨col鍒楃殑鍊煎ぇ浜?.5鐨勮
• df.sort_values(col1)锛氭寜鐓у垪col1鎺掑簭鏁版嵁锛岄粯璁ゅ崌搴忔帓鍒?/li>
• df.sort_values(col2, ascending=False)锛氭寜鐓у垪col1闄嶅簭鎺掑垪鏁版嵁
• df.sort_values([col1,col2], ascending=[True,False])锛氬厛鎸夊垪col1鍗囧簭鎺掑垪锛屽悗鎸塩ol2闄嶅簭鎺掑垪鏁版嵁
• df.groupby(col)锛氳繑鍥炰竴涓寜鍒梒ol杩涜鍒嗙粍鐨凣roupby瀵硅薄
• df.groupby([col1,col2])锛氳繑鍥炰竴涓寜澶氬垪杩涜鍒嗙粍鐨凣roupby瀵硅薄
• df.groupby(col1)[col2]锛氳繑鍥炴寜鍒梒ol1杩涜鍒嗙粍鍚庯紝鍒梒ol2鐨勫潎鍊?/li>
• df.pivot_table(index=col1, values=[col2,col3], aggfunc=max)锛氬垱寤轰竴涓寜鍒梒ol1杩涜鍒嗙粍锛屽苟璁＄畻col2鍜宑ol3鐨勬渶澶у€肩殑鏁版嵁閫忚琛?/li>
• df.groupby(col1).agg(np.mean)锛氳繑鍥炴寜鍒梒ol1鍒嗙粍鐨勬墍鏈夊垪鐨勫潎鍊?/li>
• data.apply(np.mean)锛氬DataFrame涓殑姣忎竴鍒楀簲鐢ㄥ嚱鏁皀p.mean
• data.apply(np.max,axis=1)锛氬DataFrame涓殑姣忎竴琛屽簲鐢ㄥ嚱鏁皀p.max

鏁版嵁鍚堝苟

• df1.append(df2)锛氬皢df2涓殑琛屾坊鍔犲埌df1鐨勫熬閮?/li>
• df.concat([df1, df2],axis=1)锛氬皢df2涓殑鍒楁坊鍔犲埌df1鐨勫熬閮?/li>
• df1.join(df2,on=col1,how=鈥榠nner鈥?锛氬df1鐨勫垪鍜宒f2鐨勫垪鎵цSQL褰㈠紡鐨刯oin

鏁版嵁缁熻

• df.describe()锛氭煡鐪嬫暟鎹€煎垪鐨勬眹鎬荤粺璁?/li>
• df.mean()锛氳繑鍥炴墍鏈夊垪鐨勫潎鍊?/li>
• df.corr()锛氳繑鍥炲垪涓庡垪涔嬮棿鐨勭浉鍏崇郴鏁?/li>
• df.count()锛氳繑鍥炴瘡涓€鍒椾腑鐨勯潪绌哄€肩殑涓暟
• df.max()锛氳繑鍥炴瘡涓€鍒楃殑鏈€澶у€?/li>
• df.min()锛氳繑鍥炴瘡涓€鍒楃殑鏈€灏忓€?/li>
• df.median()锛氳繑鍥炴瘡涓€鍒楃殑涓綅鏁?/li>
• df.std()锛氳繑鍥炴瘡涓€鍒楃殑鏍囧噯宸?/li>