pandas删除某列有空值的行pandas中处理缺失值dropna

pandas删除某列有空值的行_pandas中处理缺失值：dropna分析数据免不了遇到很多空值的情况，如果想去除这些空值，pandas设置了专门的函数：dropna(),下面将对dropna()进行详细的介绍dropna()pandas.DataFrame.dropnaDataReaps.dropna,extsrO,Mtv-制J/1,irHS/i-Vofle,subsNone,旳跑碎Fst：se)Removemissingvau&5.Seetfi&UserGlideformoreonwnidivak帖e怅consideredmissing；aridhowtowork戒hrrissindata.ss:0ar也他(；JarcoinsoetsuRODel&imineifrsw?w：3)umnswhicTcoitanmissrigalliesareremove-d.0Horincex:Dropm/swnichcontainmusing旧山es,1,otcolimr$13rcpcolJims时1出sortainniissngvalie.De閑eca曲s-nw艸洁划GE3,D:Passtupleorlistlodropcnmultifileaxes.On、ssng：eisabw阳也how:刮冀股址出DelermineifrawercolumnislemovedMmDalaFrame,MrfienwehaealI住asEoneParaniQtQrs:NAor胡hALan/:FanvNAvHues日飞present,drolhalnowmeolumn.3Uallvaluesa阳NidropIhaErtwortx-limn.ttiresh:皿UpHOK/F.eiirethatfllStiyfOfl-NAvfllEW-需要重点掌握的知识点：LabelsaluigolheitetoconsicEr,e.c.tyuarecmppingrtwrsliiesewJdb&a第一点需要确定的参数就是axis,O:行，1:列当inplace=True时，how建议设置为allIfTrtiidft卄航白严也卅“1齐旳Jo广口！rt14ritiirn气片血亠建议采用默认返回新对象的方法，不要对原始数据进行修改subset建议每次都用上，更有针对性thresh为非空的值得数量，小于该数量将会被删除首先需要判断是否含有空值:isna()df.isna()结果nametoyborn0FalseTrueTrue1FalseFalseFalseisnull()FalseTruedf.isnull()结果nametoybornDFalseTrueTrue1FalseFalseFalse判断是否全部为空：FalseTrueisna().any()orisnull.any(),两个函数是一样的df.isnull().any()结果：nameFalsetoyTruebornTrue判断某一列是否为空：booldftoy.isnull()dftoy.isnull().any()dftoy1isnullexecutedin18ms,finished17:38:072019-08-240True1 False2 FalseName:toyrdtype:booldf1toy.isnull()any|executedin13ms,finished17:38:092019-08-24F面正式学习：dropna()DataFrame.dropna(self,axis=0,how=any,thresh=None,subset=None,inplace=False)sourceaxis:为轴方向：默认为axis=0当axis=0,当某行出现缺失值时，将该行丢弃并返回当axis=1,当某列出现缺失值时，将改列丢弃并返回how:确定缺失值的个数：缺省时为how=anyhow=any，表明只要某行或者列出现缺失值就将该行列丢弃how=a，表明某行列全部为缺失值才将其丢弃thresh:阈值设定当行列中非缺省值的数量少于给定的值就将该行丢弃subset:部分标签中删除某行列subset=a,d即丢弃子列ad中含有缺失值的行iniplace:bool取值，默认False当inplace=True,即对原数据操作，没有返回值实例学习:pd.dropna():df=pd.DataFrame(name:Alfred,Batman,Catwoman,toy:np.nan,Batmobile,Bullwhip,born:pd.NaT,pd.Timestamp(1940-04-25),pd.NaT)rs=df.dropna()print(df)print(=*40)print(rs)默认设置情况下:结果:nametoyborn0AlfredNaN1BatmanEatinobile1940-04-252CatwomanBullwhipNaTnametoybornaxis:1BatmanBatmobile1940-04-25默认为0,删除含有缺失值的行，axis=1删除含有缺失值的列rs=df.dropna(axis=1)结果:nametoyborn0AlfredNaNNaT1BatmanBatmobile1940亠*252CatwomanBullwhipNaTname0Alfredhow:默认为any1Batmanhow=any，表明只要某行或者列出现缺失值就将该行列丢弃how=a，表明某行列全部为缺失值才将其丢弃重新构建数据，增加一列和一行空值：df=pd.DataFrame(name:Alfred,Batman,Catwoman,toy:np.nan,Batmobile,Bullwhip,born:pd.NaT,pd.Timestamp(1940-04-25),pd.NaT)dfp=np.nandf.loc4=np.nanhow=all:axis=1:rs=df.dropna(axis=1,how=all)结果：nametoybornP0AlfredNaNNaT购N1BatmanBatmobile1940-04-25脱N2CatwomanBullwhipNaTNaN4NaNNaNNaTKaNnametoyborn0AlfredNaNNaT1BatnanBatmobile1940-04-25axis=0:-CatwomanBullwhipNaTrs=df.dropna(axis=0,how=all)结果nametoybornP0AlfredNaNNaTKaN1BatmanBatmobile1540-04-25KaN2CatwomanBullwhipNa?NaN4NaNNaNNa?KaNnajnetoyboraP0AlfredNaNNa?KaNthresh:BatmanBatmobile1940-04-25NaN行或列至少保留的非空值的数量，关键是非空的数量NaM传入一个整数值，当行或列低于该值时删除，大于等于时不删除全部为空时删除当没行至少有一个不是空值时保留，rs=df.dropna(axis=0,thresh=1)name结果bornAlfredWaTNaNBatir.anBatmobile1940-04-25NaNCatwoarLanBullwhipNaTNaNNaNNaNNaTMaNnametoyborn0AlfredKaN当每行至少有2个不是空值时保留，全部为空时删除.NaTNaNNaNrs=df.dropna(axis=O,thresh=2)i:结果nametoybornP0AlfredNaNNaTNatl1BatmanBatmobile1940-04-25NaN2CatwomanBullwhipNaTNaN4NaNNaNNaTNaNnametoybornpsubset:注意，只能删除行，需要给定列标签，不能删除列subset=加即丢弃襦列ad中含有缺失值的行NflT1NaN删除toy中含有空值的行，rs=df.dropna(axis=0,subset=toy)结果nametoybornPUAlfredNaNKaTNaN1BatmanBatmobile1940-04-25NaK2CatwomanBullwhipKaTNaN4NaNNaNNaTNaKnametoybornFinjpiace：Batntaj：Bal:r：obilDLi140-04-25默认返回新的对象，如果需要对原始数据进行修改，可以设置为:True=print(df)print(=*40)df=pd.DataFrame(name:Aifred,Batman,Catwoman,toy:np.nan,Batmobiie,Buiiwhip,born:pd.NaT,pd.Timestamp(1940-04-25),pd.NaT)dfp=np.nandf.ioc4=np.nan#rs=df.dropna(axis=0,subset=toy)df.dropna(axis=0,subset=toy,inpiace=True)print(df)结果nametoybornP0AlfredNaNNaTNaN1BatmanBatmobile1940-04-25NaN2CatwomanBullwhipNaTNaN4NaNNaNNaTNaNnametoybornP1BatmanBatmobile1940-04-25NaN推荐学习链接：英文版解释:DataFrame.dropna(self,axis=0,how=any,thresh=None,subset=None,inplace=False)sourceRemovemissingvalues.SeetheUserGuideformoreonwhichvaluesareconsideredmissing,andhowtoworkwithmissingdata.Parameters:axis:0orindex,1orcolumns,default0Determineifrowsorcolumnswhichcontainmissingvaluesareremoved.0,orindex:Droprowswhichcontainmissingvalues.1,orcolumns:Dropcolumnswhichcontainmissingvalue.Deprecatedsinceversion0.23.0:Passtupleorlisttodroponmultipleaxes.Onlyasingleaxisisallowed.how:any,all,defaultanyDetermineifroworcolumnisremovedfromDataFrame,whenwehaveatleastoneNAorallNA.any:fanyNAvaluesarepresent,dropthatroworcolumn.all:IfallvaluesareNA,dropthatroworcolumn.thresh:int,optionalRequirethatmanynon-NAvalues.subset:array-like,optionalLabelsalongotheraxistoconsider,e.g.ifyouaredroppingrowsthesewouldbealistofcolumnstoinclude.inplace:bool,defaultFalseIfTrue,dooperationinplaceandreturnNone.