KeyError Pandas – How To Fix

Pandas KeyError is frustrating. This mistake happens because Pandas cannot find what you lot're looking for.

To fix this either:

  1. Preferred Option: Make sure that your column label (or row label) is in your dataframe!
  2. Error catch option: Utilise df.get('your cavalcade') to look for your cavalcade value. No error will be thrown if it is non found.
                ane. df.go('your_column', default=value_if_no_column)

Pseudo code: Check to see if a column is in your dataframe, if non, return the default value.

Pandas KeyError

In most cases, think of 'key' as the aforementioned as 'proper name.' Pandas is telling you that information technology can not find your column name. The preferred method is to *make sure your column name is in your dataframe.*

OR if you want to try and catch your mistake, you tin can use df.go('your_column'). However, if yous don't know what columns are in your dataframe…practice you really know your data?

It's best to head back upstream with your code and debug where your expectations and dataframe columns mismatch.

Endeavour, Except

For a general solution, you lot can use the Attempt Except convention to catch errors in your code. Withal, beware. Using a coating Endeavor/Except clause is dangerous and poor code practice if you lot do non know what you are doing.

If you are 'communicable' full general errors with try/except, this ways that anything can slip through your code. This could result in unexpected errors getting through and a spider web of complexity.

Goal: Try to never allow the reality of your code go too far away from the expectations of your lawmaking.

Let'due south accept a look at a sample:

Pandas KeyError¶

Pandas KeyError can exist annoying. It by and large happens when pandas cannot notice the affair yous're looking for. Commonly this is to due a column information technology cannot find. It'south simple to debug!

Let'due south check out some examples:

  1. Locating the error
  2. Fixing the error via the root crusade
  3. Catching the error with df.get()

Commencement, permit's create a DataFrame

In [5]:

                                            df                      =                      pd                      .                      DataFrame                      ([(                      'Foreign Picture palace'                      ,                      'Restaurant'                      ),                      (                      'Liho Liho'                      ,                      'Restaurant'                      ),                      (                      '500 Club'                      ,                      'bar'                      ),                      (                      'The Square'                      ,                      'bar'                      )],                      columns                      =                      (                      'name'                      ,                      'blazon'                      )                      )                      df

Out[5]:

proper name type
0 Strange Movie theater Eating house
i Liho Liho Restaurant
2 500 Social club bar
3 The Foursquare bar

Now permit's try to call a column that is in our dataframe and is Not in our dataframe

In [6]:

                                            # Is in our dataframe                      df                      [                      'name'                      ]

Out[6]:

0    Foreign Movie theater 1         Liho Liho ii          500 Club 3        The Square Name: name, dtype: object

In [9]:

                                            # Is non in our dataframe                      df                      [                      'nutrient'                      ]                    
                    ---------------------------------------------------------------------------                    KeyError                    Traceback (most contempo telephone call last)                    /opt/anaconda3/lib/python3.seven/site-packages/pandas/cadre/indexes/base.py                    in                    get_loc                    (self, key, method, tolerance)                                          2888                    endeavor                    :                    -> 2889                                                            return                    self._engine.get_loc(casted_key)                                          2890                    except                    KeyError                    as                    err:                    pandas/_libs/index.pyx                    in                    pandas._libs.index.IndexEngine.get_loc                    ()                    pandas/_libs/index.pyx                    in                    pandas._libs.index.IndexEngine.get_loc                    ()                    pandas/_libs/hashtable_class_helper.pxi                    in                    pandas._libs.hashtable.PyObjectHashTable.get_item                    ()                    pandas/_libs/hashtable_class_helper.pxi                    in                    pandas._libs.hashtable.PyObjectHashTable.get_item                    ()                    KeyError: 'nutrient'  The above exception was the straight cause of the following exception:                    KeyError                    Traceback (near recent telephone call last)                    <ipython-input-9-dce7ca21d87e>                    in                    <module>                                          1                    # Is not in our dataframe                    ----> 2                                        df[                    'food'                    ]                    /opt/anaconda3/lib/python3.7/site-packages/pandas/cadre/frame.py                    in                    __getitem__                    (self, key)                                          2897                    if                    self.columns.nlevels                    >                    ane                    :                                          2898                    return                    self._getitem_multilevel(cardinal)                    -> 2899                                        indexer                    =                    cocky.columns.get_loc(key)                                          2900                    if                    is_integer(indexer)                    :                                          2901                    indexer                    =                    [indexer]                    /opt/anaconda3/lib/python3.7/site-packages/pandas/core/indexes/base of operations.py                    in                    get_loc                    (cocky, central, method, tolerance)                                          2889                    return                    self._engine.get_loc(casted_key)                                          2890                    except                    KeyError                    as                    err:                    -> 2891                                                            heighten                    KeyError(key)                    from                    err                                          2892                                          2893                    if                    tolerance                    is                    not                    None                    :                    KeyError: 'nutrient'

Oh no! We got a KeyError. This means that Pandas cannot find "food" within our dataframe. We know why this is. Simply, it's not in our DF.

To become effectually this, either add together a 'food' column. Or use df.get() to try and take hold of it. Hither nosotros will use .get() and notice there is no error thrown.

This mistake can also happen when yous're trying to admission an index (for rows) label that doesn't exist. Check out this instance

Out[22]:

name    500 Guild blazon         bar Name: 2, dtype: object
                  ---------------------------------------------------------------------------                  ValueError                  Traceback (most contempo call last)                  /opt/anaconda3/lib/python3.7/site-packages/pandas/core/indexes/range.py                  in                  get_loc                  (self, key, method, tolerance)                                      350                  try                  :                  --> 351                                                      render                  self._range.index(new_key)                                      352                  except                  ValueError                  as                  err:                  ValueError: nine is not in range The above exception was the directly crusade of the following exception:                  KeyError                  Traceback (most recent call final)                  <ipython-input-23-026debeea1cf>                  in                  <module>                                      one                  # Does not                  ----> 2                                    df.loc[                  9                  ]                  /opt/anaconda3/lib/python3.7/site-packages/pandas/core/indexing.py                  in                  __getitem__                  (self, fundamental)                                      877                                      878                  maybe_callable                  =                  com.apply_if_callable(key,                  cocky.obj)                  --> 879                                                      return                  self._getitem_axis(maybe_callable,                  axis=centrality)                                      880                                      881                  def                  _is_scalar_access(self,                  key:                  Tuple)                  :                  /opt/anaconda3/lib/python3.7/site-packages/pandas/core/indexing.py                  in                  _getitem_axis                  (self, central, axis)                                      1108                  # fall thru to straight lookup                                      1109                  self._validate_key(key,                  centrality)                  -> 1110                                                      return                  cocky._get_label(key,                  axis=axis)                                      1111                                      1112                  def                  _get_slice_axis(cocky,                  slice_obj:                  slice,                  centrality:                  int)                  :                  /opt/anaconda3/lib/python3.seven/site-packages/pandas/cadre/indexing.py                  in                  _get_label                  (self, characterization, axis)                                      1057                  def                  _get_label(self,                  characterization,                  axis:                  int)                  :                                      1058                  # GH#5667 this will fail if the characterization is non present in the axis.                  -> 1059                                                      render                  self.obj.xs(characterization,                  axis=axis)                                      1060                                      1061                  def                  _handle_lowerdim_multi_index_axis0(self,                  tup:                  Tuple)                  :                  /opt/anaconda3/lib/python3.vii/site-packages/pandas/core/generic.py                  in                  xs                  (self, key, centrality, level, drop_level)                                      3480                  loc,                  new_index                  =                  cocky.index.get_loc_level(cardinal,                  drop_level=drop_level)                                      3481                  else                  :                  -> 3482                                    loc                  =                  self.index.get_loc(central)                                      3483                                      3484                  if                  isinstance(loc,                  np.ndarray)                  :                  /opt/anaconda3/lib/python3.seven/site-packages/pandas/core/indexes/range.py                  in                  get_loc                  (cocky, key, method, tolerance)                                      351                  return                  self._range.index(new_key)                                      352                  except                  ValueError                  as                  err:                  --> 353                                                      heighten                  KeyError(key)                  from                  err                                      354                  enhance                  KeyError(key)                                      355                  return                  super(                  )                  .get_loc(key,                  method=method,                  tolerance=tolerance)                  KeyError: ix

This example immediately jumps out to me and says "hey, I tin can not observe the characterization 9 in your row. Do something well-nigh this"

If you wanted to programatically do this, the long way would be to check for it first. If it is in the index, so proceed. But ordinarily, I want to brand sure I'chiliad calling something I KNOW is in my index.

In [25]:

                                        value                    =                    9                    if                    value                    in                    df                    .                    index                    :                    impress                    (                    df                    .                    loc                    [                    value                    ])                    else                    :                    print                    (                    "Not in alphabetize"                    )

In [26]:

                                            value                      =                      2                      if                      value                      in                      df                      .                      alphabetize                      :                      print                      (                      df                      .                      loc                      [                      value                      ])                      else                      :                      print                      (                      "Not in index"                      )                    
proper name    500 Lodge type         bar Name: 2, dtype: object

Cheque out more than Pandas functions on our Pandas Page