'e-ddlmZddlmZmZ ddlmZmZmZm Z m Z m Z m Z m Z  ddlmZddlmZdd Zddd d efd ZdZefdZddZy#e$rZedeedddZ[wwxYw#e$rGdd ZeZYMwxYw))Table) Expressionfield) DeclarationExecNodeOptionsTableSourceNodeOptionsFilterNodeOptionsProjectNodeOptionsAggregateNodeOptionsOrderByNodeOptionsHashJoinNodeOptionsz@The pyarrow installation is not built with support for 'acero' ()N)ScanNodeOptionsc4eZdZGddZGddZy)DatasetModuleStubc eZdZy)DatasetModuleStub.DatasetN__name__ __module__ __qualname__-lib/python3.11/site-packages/pyarrow/acero.pyDatasetr2 rrc eZdZy)!DatasetModuleStub.InMemoryDatasetNrrrrInMemoryDatasetr5rrrN)rrrrrrrrrr1s    rrTc xtdt||}|jjDcgc] }t |}}tj |tdt |g}|jjd}|*tj |tdt|g}|Scc}w)Nscan use_threadsprojectfilter) rrschemanamesr from_sequencer _scan_optionsgetr )datasetr#declf projections filter_exprs r_dataset_to_declr0:s vwKP QD&-^^%9%9:58:K:  $ $ {9&8&EFG D''++H5K(( ;x):;)GH I  K;sB7Fc t|ttjfst dt |t|ttjfst dt |i} t|t tfs|g}t|D] \} } | | | < i} t|tt fs|g}t|D] \} } | | | < |jj}|jj}|dk(s|dk(rg}nE|dk(s|dk(rg}n8|dk(s|dk(r|Dcgc] }|| vs| }}n|dk(r|Dcgc] }|| vs| }}i}t|D]\} }||vs | ||<i}t|D]\} }||vs | ||<t|tjrt|| }ntd t|}t|tjrt|| }ntd t|}|rt||||||xsd |xsd }nt||||xsd |xsd }td |||g}|rL|dk(rFt|}t|}t!|}g}g}t||zD]\} }| t!|krq||vrm|j#|||| |}|j#t%j&dt%j(| t%j(||zg| |k\r||vr|r| |kr ||vr||z }|r| |k\r ||vr||z }|j#||j#t%j(| tdt+||}tj,||g}|j/| }| tk(r|S| tj0k(rtj0|St dcc}wcc}w)a\ Perform join of two tables or datasets. The result will be an output table with the result of the join operation Parameters ---------- join_type : str One of supported join types. left_operand : Table or Dataset The left operand for the join operation. left_keys : str or list[str] The left key (or keys) on which the join operation should be performed. right_operand : Table or Dataset The right operand for the join operation. right_keys : str or list[str] The right key (or keys) on which the join operation should be performed. left_suffix : str, default None Which suffix to add to left column names. This prevents confusion when the columns in left and right operands have colliding names. right_suffix : str, default None Which suffix to add to the right column names. This prevents confusion when the columns in left and right operands have colliding names. use_threads : bool, default True Whether to use multithreading or not. coalesce_keys : bool, default False If the duplicated keys should be omitted from one of the sides in the join result. output_type: Table or InMemoryDataset The output type for the exec plan result. Returns ------- result_table : Table or InMemoryDataset zExpected Table or Dataset, got z left semiz left antiz right semiz right antiinnerz left outerz right outerr" table_source)output_suffix_for_leftoutput_suffix_for_righthashjoin)optionsinputsz full outercoalescer$Unsupported output type) isinstancerdsr TypeErrortypetuplelist enumerater&r'r0rrr setlenappendr_call_fieldr r(to_tabler) join_type left_operand left_keys right_operand right_keys left_suffix right_suffixr# coalesce_keys output_typeleft_keys_orderidxkeyright_keys_order left_columns right_columnscolleft_column_keys_indicescolnameright_column_keys_indices left_source right_source join_optsr,left_columns_setright_columns_setright_operand_indexprojected_col_namesr.right_key_index projection result_tables r _perform_joinrfOsP lUBJJ$7 89$|:L9MNOO meRZZ%8 99$}:M9NOPPO i% /K i(#S"# j4- 0 \ j)$S #$ &&,,L!((..MK9 #; l "i<&? g l!:( C7G,GC   m #' 3o+EC  "!,/4 W i 03 $W -4!#!-05 W j 14 %g .5 , +&|M !.2H2VW -,' ;O " 2=A ' y*lM#.#4"$0$6B ( y*#.#4"$0$6B I{L.I Dl2|, .!,/  !,">?$ HCS&&32J+J#**3/ #<s34#6""$$Z"))#. 0A0A/?1A2 ++7P0P  C*=$=00;&C S,?%?//<'C#**3/""%%c*E$ J! )+7JK (($ );<==[=9Le ** *!!,//122U  s) O/3O/ O4 O4c tjtdt|tdt|g}|j dS)aFilter rows of a table based on the provided expression. The result will be an output table with only the rows matching the provided expression. Parameters ---------- table : Table or Dataset Table or Dataset that should be filtered. expression : Expression The expression on which rows should be filtered. Returns ------- Table r3)r8r%Tr")rr(rr rH)table expressionr,s r _filter_tablerjsL"  $ $N,B5,IJH&7 &CD& D ==T= **rc vt|tjrt|d}nt dt |}t dt |fi|}t j||g}|jd}|tk(r|S|tjk(rtj|Std)NTr"r3order_byr;) r<r=rr0rrr r(rHrrr>)table_or_dataset sort_keysrQkwargs data_sourcerlr,res r _sort_sourcerqs"BJJ/&'7TJ ! 23CD :'9)'Nv'NOH  $ $k8%< =D==T=2Le ** *!!,//122rc tjtdt|tdt||g}|j |S)Nr3 aggregate)keysr")rr(rr rH)rh aggregatesrtr#r,s r _group_byrv.sL  $ $N$:5$ABK!5jt!LM& D ==[= 11r)T) pyarrow.librpyarrow.computerrpyarrow._acerorrrr r r r r ImportErrorexcstrpyarrow.datasetr+r=pyarrow._datasetrrr0rfrjrqrvrrrrs2-     0.#"%# n3b+0;@3,2O  J3s8*TUV   Bs(A A+ A(A##A(+A?>A?