U
|
ý°d¡" ã
|
@ sº d Z ddlmZ ddlZddlZddlZddlmZ ddl Z
|
ddl m Z m Z m Z ddlmZ ej d¡Zej d¡Zeej dd e d d e d
|
g¡d d Zeej dd d idd ig¡dd Zedd Zedd Zeejjdd Zdd Z edd Z edd Z d d Z ej d d d!g¡ej d"d#d d$gd#d$d gd#d d$d%gd$d d#gg¡d&d' Z d(d) Z!d*d+ Z"ed,d- Z#d.d/ Z$d0d1 Z%ej d2d3d4d5d6g¡d7d8 Z&d9d: Z'dS );zW
|
Tests dtype specification during parsing
|
for all of the parsers defined in parsers.py
|
é )ÚStringION)ÚCategoricalDtype)Ú CategoricalÚ DataFrameÚ TimestampZ pyarrow_xfailZ pyarrow_skipÚdtypeÚcategory©ÚaÚbÚcc C sX | }d}t tdddgtdddgtdddgd}|jt||d }t ||¡ d S )
|
Nú a,b,c
|
1,a,3.4
|
1,a,3.4
|
2,b,4.5Ú1Ú2r
|
r ú3.4ú4.5r ©r ©r r Úread_csvr ÚtmÚassert_frame_equal©Ú all_parsersr ÚparserÚdataÚexpectedÚactual© r úed:\z\workplace\vscode\pyvenv\venv\Lib\site-packages\pandas/tests/io/parser/dtypes/test_categorical.pyÚtest_categorical_dtype s ýÿr r é c C sP | }d}t dddgtdddgdddgd}|jt||d }t ||¡ d S )
|
Nr r é r
|
r g333333 @g @r r r r r r r Ú test_categorical_dtype_single4 s ÿr" c C sX | }d}t tdddgtdddgtdddgd}|jt|d d
|
}t ||¡ d S ) Nz a,b,c
|
1,b,3.4
|
1,b,3.4
|
2,a,4.5r r r r
|
r r r r r r ©r r r r r r r r Ú test_categorical_dtype_unsortedD s ýÿr$ c C sZ | }d}t tdddgtdtjdgtdddgd}|jt|d d
|
}t ||¡ d S ) Nz a,b,c
|
1,b,3.4
|
1,nan,3.4
|
2,a,4.5r r r r
|
r r r r r )r r ÚnpÚnanr r r r r# r r r Ú test_categorical_dtype_missingW s ýÿr' c C s | }t dd tdD ¡}tdt|ddi}|jtdd |¡ d d
|
}|d jj t |j
|
jj ¡dd|d< t ||¡ d S ) Nc S s g | ] }t |qS r )Ústr)Ú.0Úir r r Ú
|
<listcomp>o s zCtest_categorical_dtype_high_cardinality_numeric.<locals>.<listcomp>i r
|
T)Úorderedza
|
Ú
|
r r )r% ÚsortÚranger r r r ÚjoinÚcatZreorder_categoriesr
|
Ú
|
categoriesr r r# r r r Ú/test_categorical_dtype_high_cardinality_numericj s
|
ÿ
|
r3 c C sV t j |d¡}| }d}d}|j|||d}| t¡}|j|||dd}t ||¡ d S )Nz utf16_ex.txtzutf-16ú )ÚsepÚencodingr )r5 r6 r )ÚosÚpathr0 r Úapplyr r r )r Ú csv_dir_pathÚpthr r6 r5 r r r r r Ú test_categorical_dtype_utf16y s
|
r< c C s | }d}t ddgtddgdt ddgtddgdddgd g}|jt|dd
|
idd &}t||D ]\}}t ||¡ qhW 5 Q R X d S ) Núa,b
|
1,a
|
1,b
|
1,b
|
2,cr r
|
r ©r
|
r r! r é ©Úindexr ©r Ú chunksize)r r r r Úzipr r )r r r Ú expectedsÚactualsr r r r r Ú1test_categorical_dtype_chunksize_infer_categories s þ ÿrG c C s¦ | }d}dddg}t ddgtddg|ddt ddgtddg|dddd gd
|
g}t|}|jt|d|idd &}t||D ]\}}t ||¡ qW 5 Q R X d S ) Nr= r
|
r r r ©r2 r> r! r? r@ rB )r r r r r rD r r ) r r r ZcatsrE r rF r r r r r Ú4test_categorical_dtype_chunksize_explicit_categories s
|
þþrI c C s\ t j |d¡}| }d}|j|d |d}t|d |d< |j|d |ddid}t ||¡ d S )Nzunicode_series.csvzlatin-1)Úheaderr6 r r )rJ r6 r )r7 r8 r0 r r r r )r r: r; r r6 r r r r r Ú test_categorical_dtype_latin1² s rK r, FTr2 r
|
r Údc C sb | }d}t ddddgtddddg||dd}dt||di}|jt||d }t ||¡ d S )
|
Nr= r r! r
|
r r )r2 r, r> r )r r r r r r r )r r2 r, r r r r Úresultr r r Ú test_categorical_category_dtype¿ s
|
|
ÿþÿ rN c C sh | }d}t dddg}tddddgtddddgdddgdd}|jt|d|id }t ||¡ d S )
|
Nr= r r r
|
r r! rH r> r ©r r r r r r r )r r r r r rM r r r Ú(test_categorical_category_dtype_unsortedÙ s
|
þÿrP c C sT | }dt dddgi}d}tdtddddgi}|jt||d}t ||¡ d S )Nr r r! r? z b
|
1
|
1
|
2
|
3r rO ©r r r r r rM r r r Ú test_categorical_coerces_numericì s rR c C s` | }t jdddgd d}dt|i}d}tdt|d ji}|jt||d}t ||¡ d S )Nz
|
2017-01-01z
|
2018-01-01z
|
2019-01-01)Úfreqr z"b
|
2017-01-01
|
2018-01-01
|
2019-01-01r )
|
ÚpdZ DatetimeIndexr r r r2 r r r r )r r Zdtir r r rM r r r Ú!test_categorical_coerces_datetime÷ s rU c C sV | }dt tdgi}d}tdttdgd i}|jt||d}t ||¡ d S )Nr Z2014zb
|
2014-01-01
|
2014-01-01r! r )r r r r r r r r rQ r r r Ú"test_categorical_coerces_timestamp s rV c C sX | }dt t dddg¡i}d}tdt|d ji}|jt||d}t ||¡ d S )Nr Z1HZ2HZ3Hz
|
b
|
1H
|
2H
|
3Hr )
|
r rT Z to_timedeltar r r2 r r r r rQ r r r Ú"test_categorical_coerces_timedelta s rW r zb
|
True
|
False
|
NA
|
Falsezb
|
true
|
false
|
NA
|
falsezb
|
TRUE
|
FALSE
|
NA
|
FALSEzb
|
True
|
False
|
NA
|
FALSEc C sN | }dt ddgi}tdtddd dgi}|jt||d}t ||¡ d S )Nr FTr rO )r r r r r rM r r r Ú&test_categorical_dtype_coerces_boolean s
|
rX c C sZ | }dt ddddgi}d}tdttd|d di}|jt||d}t ||¡ d S )Nr r
|
rL Úez b
|
d
|
a
|
c
|
dZdacdr )r r r Úlistr r r r rQ r r r Ú&test_categorical_unexpected_categories- s r[ )(Ú__doc__Úior r7 Únumpyr% ZpytestZpandas.core.dtypes.dtypesr ZpandasrT r r r Zpandas._testingZ_testingr ÚmarkZ usefixturesZ xfail_pyarrowZ skip_pyarrowZ parametrizer r" r$ r' Zslowr3 r< rG rI rK rN rP rR rU rV rW rX r[ r r r r Ú<module> sr ýþ
|
|
|
|
$þ
|
üþ
|
