pandas小记：pandas基本设置

小咪咪 2022-08-09 15:53 301阅读 0赞

http://[blog.csdn.net/pipisorry/article/details/49519545][blog.csdn.net_pipisorry_article_details_49519545]

## 安装 ##

Bugfix  
from pandas.core.groupby.groupby import DataFrameGroupBy, \\ E ImportError: cannot import name 'DataFrameGroupBy'  
解决：  
pip install tqdm --upgrade  
\[[Pandas not working: DataFrameGroupBy ; PanelGroupBy][Pandas not working_ DataFrameGroupBy _ PanelGroupBy]\]

# pandas输出格式设置示例 #

## pandas输出精度局部设置local setting ##

with pd.option_context('display.precision', 3):
        print(df)

**Note**: 试了好久终于找到了这种设置方法！

**它是这样实现的**

class option_context(object):
        """
        >>> with option_context('display.max_rows', 10, 'display.max_columns', 5):
    
        """
    
        def __init__(self, *args):
            if not (len(args) % 2 == 0 and len(args) >= 2):
                raise ValueError(
                    'Need to invoke as'
                    'option_context(pat, val, [(pat, val), ...)).'
                )
    
            self.ops = list(zip(args[::2], args[1::2]))
    
        def __enter__(self):
            undo = []
            for pat, val in self.ops:
                undo.append((pat, _get_option(pat, silent=True)))
    
            self.undo = undo
    
            for pat, val in self.ops:
                _set_option(pat, val, silent=True)
    
        def __exit__(self, *args):
            if self.undo:
                for pat, val in self.undo:
                    _set_option(pat, val, silent=True)

其实类似numpy数组输出精度的设置：\[[numpy输入输出][numpy]\]

## pandas浮点数据输出禁用科学计数法的方式 ##

with pd.option_context('display.float_format', lambda x: '%.3f' % x)

当然series也可以这样

Series(np.random.randn(3)).apply(lambda x: '%.3f' % x)

[皮皮blog][blog]

# pandas选项和设置 #

## List of Options ##

<table> 
 <thead> 
  <tr> 
   <th>Option</th> 
   <th>Default</th> 
   <th>Function</th> 
  </tr> 
 </thead> 
 <tbody> 
  <tr> 
   <td>display.chop_threshold</td> 
   <td>None</td> 
   <td>If set to a float value, all floatvalues smaller then the giventhreshold will be displayed asexactly 0 by repr and friends.</td> 
  </tr> 
  <tr> 
   <td>display.colheader_justify</td> 
   <td>right</td> 
   <td>Controls the justification ofcolumn headers. used by DataFrameFormatter.</td> 
  </tr> 
  <tr> 
   <td>display.column_space</td> 
   <td>12</td> 
   <td>No description available.</td> 
  </tr> 
  <tr> 
   <td>display.date_dayfirst</td> 
   <td>False</td> 
   <td>When True, prints and parses dateswith the day first, eg 20/01/2005</td> 
  </tr> 
  <tr> 
   <td>display.date_yearfirst</td> 
   <td>False</td> 
   <td>When True, prints and parses dateswith the year first, eg 2005/01/20</td> 
  </tr> 
  <tr> 
   <td>display.encoding</td> 
   <td>UTF-8</td> 
   <td>Defaults to the detected encodingof the console. Specifies the encodingto be used for strings returned byto_string, these are generally stringsmeant to be displayed on the console.</td> 
  </tr> 
  <tr> 
   <td>display.expand_frame_repr</td> 
   <td>True</td> 
   <td>Whether to print out the full DataFramerepr for wide DataFrames acrossmultiple lines,max_columns isstill respected, but the output willwrap-around across multiple “pages”if its width exceedsdisplay.width.</td> 
  </tr> 
  <tr> 
   <td><strong>display.float_format</strong></td> 
   <td>None</td> 
   <td>The callable should accept a floatingpoint number and return a string withthe desired format of the number.This is used in some places likeSeriesFormatter.See core.format.EngFormatter for an example.</td> 
  </tr> 
  <tr> 
   <td>display.height</td> 
   <td>60</td> 
   <td>Deprecated. Use display.max_rows instead.</td> 
  </tr> 
  <tr> 
   <td>display.large_repr</td> 
   <td>truncate</td> 
   <td>For DataFrames exceeding max_rows/max_cols,the repr (and HTML repr) can showa truncated table (the default from 0.13),or switch to the view from df.info()(the behaviour in earlier versions of pandas).allowable settings, [‘truncate’, ‘info’]</td> 
  </tr> 
  <tr> 
   <td>display.line_width</td> 
   <td>80</td> 
   <td>Deprecated. Use display.width instead.</td> 
  </tr> 
  <tr> 
   <td>display.max_columns</td> 
   <td>20</td> 
   <td>max_rows and max_columns are usedin __repr__() methods to decide ifto_string() or info() is used torender an object to a string. Incase python/IPython is running ina terminal this can be set to 0 andpandas will correctly auto-detectthe width the terminal and swap toa smaller format in case all columnswould not fit vertically. The IPythonnotebook, IPython qtconsole, or IDLEdo not run in a terminal and henceit is not possible to do correctauto-detection. ‘None’ value meansunlimited.</td> 
  </tr> 
  <tr> 
   <td>display.max_colwidth</td> 
   <td>50</td> 
   <td>The maximum width in characters ofa column in the repr of a pandasdata structure. When the column overflows,a ”...” placeholder is embedded inthe output.</td> 
  </tr> 
  <tr> 
   <td>display.max_info_columns</td> 
   <td>100</td> 
   <td>max_info_columns is used in DataFrame.infomethod to decide if per column informationwill be printed.</td> 
  </tr> 
  <tr> 
   <td>display.max_info_rows</td> 
   <td>1690785</td> 
   <td>df.info() will usually show null-countsfor each column. For large framesthis can be quite slow. max_info_rowsand max_info_cols limit this nullcheck only to frames with smallerdimensions then specified.</td> 
  </tr> 
  <tr> 
   <td>display.max_rows</td> 
   <td>60</td> 
   <td>This sets the maximum number of rowspandas should output when printingout various output. For example,this value determines whether therepr() for a dataframe prints outfully or just a summary repr.‘None’ value means unlimited.</td> 
  </tr> 
  <tr> 
   <td>display.max_seq_items</td> 
   <td>100</td> 
   <td>when pretty-printing a long sequence,no more then max_seq_items willbe printed. If items are omitted,they will be denoted by the additionof ”...” to the resulting string.If set to None, the number of itemsto be printed is unlimited.</td> 
  </tr> 
  <tr> 
   <td>display.memory_usage</td> 
   <td>True</td> 
   <td>This specifies if the memory usage ofa DataFrame should be displayed when thedf.info() method is invoked.</td> 
  </tr> 
  <tr> 
   <td>display.mpl_style</td> 
   <td>None</td> 
   <td>Setting this to ‘default’ will modifythe rcParams used by matplotlibto give plots a more pleasing visualstyle by default. Setting this toNone/False restores the values totheir initial value.</td> 
  </tr> 
  <tr> 
   <td>display.multi_sparse</td> 
   <td>True</td> 
   <td>“Sparsify” MultiIndex display (don’tdisplay repeated elements in outerlevels within groups)</td> 
  </tr> 
  <tr> 
   <td>display.notebook_repr_html</td> 
   <td>True</td> 
   <td>When True, IPython notebook willuse html representation forpandas objects (if it is available).</td> 
  </tr> 
  <tr> 
   <td>display.pprint_nest_depth</td> 
   <td>3</td> 
   <td>Controls the number of nested levelsto process when pretty-printing</td> 
  </tr> 
  <tr> 
   <td><strong>display.precision</strong></td> 
   <td>6</td> 
   <td>Floating point output precision interms of number of places after thedecimal, for regular formatting as wellas scientific notation. Similar tonumpy’sprecision print option</td> 
  </tr> 
  <tr> 
   <td>display.show_dimensions</td> 
   <td>truncate</td> 
   <td>Whether to print out dimensionsat the end of DataFrame repr.If ‘truncate’ is specified, onlyprint out the dimensions if theframe is truncated (e.g. not displayall rows and/or columns)</td> 
  </tr> 
  <tr> 
   <td>display.width</td> 
   <td>80</td> 
   <td>Width of the display in characters.In case python/IPython is running ina terminal this can be set to Noneand pandas will correctly auto-detectthe width. Note that the IPython notebook,IPython qtconsole, or IDLE do not run in aterminal and hence it is not possibleto correctly detect the width.</td> 
  </tr> 
  <tr> 
   <td>io.excel.xls.writer</td> 
   <td>xlwt</td> 
   <td>The default Excel writer engine for‘xls’ files.</td> 
  </tr> 
  <tr> 
   <td>io.excel.xlsm.writer</td> 
   <td>openpyxl</td> 
   <td>The default Excel writer engine for‘xlsm’ files. Available options:‘openpyxl’ (the default).</td> 
  </tr> 
  <tr> 
   <td>io.excel.xlsx.writer</td> 
   <td>openpyxl</td> 
   <td>The default Excel writer engine for‘xlsx’ files.</td> 
  </tr> 
  <tr> 
   <td>io.hdf.default_format</td> 
   <td>None</td> 
   <td>default format writing format, ifNone, then put will default to‘fixed’ and append will default to‘table’</td> 
  </tr> 
  <tr> 
   <td>io.hdf.dropna_table</td> 
   <td>True</td> 
   <td>drop ALL nan rows when appendingto a table</td> 
  </tr> 
  <tr> 
   <td>mode.chained_assignment</td> 
   <td>warn</td> 
   <td>Raise an exception, warn, or noaction if trying to use chainedassignment, The default is warn</td> 
  </tr> 
  <tr> 
   <td>mode.sim_interactive</td> 
   <td>False</td> 
   <td>Whether to simulate interactive modefor purposes of testing</td> 
  </tr> 
  <tr> 
   <td>mode.use_inf_as_null</td> 
   <td>False</td> 
   <td>True means treat None, NaN, -INF,INF as null (old way), False meansNone and NaN are null, but INF, -INFare not null (new way).</td> 
  </tr> 
 </tbody> 
</table>

\[[Options and Settings][]\]

## pandas.set\_option ##

pandas.set\_option(*pat*,*value*) *= <pandas.core.config.CallableDynamicDoc object at 0xb559f20c>*

Sets the value of the specified option.

示例

\[[Python pandas, widen output display?][Python pandas_ widen output display]\]

## to\_csv精度设置 ##

df_data.to_csv(outfile, index=False,
                       header=False, float_format='%11.6f')

\[[Print different precision by column with pandas.DataFrame.to\_csv()][Print different precision by column with pandas.DataFrame.to_csv]\]

from:[http://blog.csdn.net/pipisorry/article/details/49519545][blog.csdn.net_pipisorry_article_details_49519545]

ref:[pandas.core.config.get\_option][pandas.core.config.get_option]

[blog.csdn.net_pipisorry_article_details_49519545]: http://blog.csdn.net/pipisorry/article/details/49519545
[Pandas not working_ DataFrameGroupBy _ PanelGroupBy]: https://stackoverflow.com/questions/59403256/pandas-not-working-dataframegroupby-panelgroupby
[numpy]: http://blog.csdn.net/pipisorry/article/details/39088003
[blog]: http://blog.csdn.net/pipisorry
[Options and Settings]: http://pandas.pydata.org/pandas-docs/stable/options.html
[Python pandas_ widen output display]: http://www.helplib.com/qa/604294
[Print different precision by column with pandas.DataFrame.to_csv]: http://www.4byte.cn/question/373702/print-different-precision-by-column-with-pandas-dataframe-to-csv.html
[pandas.core.config.get_option]: http://nullege.com/codes/search/pandas.core.config.get_option