python-datascience/03.02-Data-Indexing-and-Selection.ipynb

{"cells":[{"metadata":{},"cell_type":"markdown","source":"# Data Indexing and Selection"},{"metadata":{},"cell_type":"markdown","source":"In [Chapter 2](02.00-Introduction-to-NumPy.ipynb), we looked in detail at methods and tools to access, set, and modify values in NumPy arrays.\nThese included indexing (e.g., ``arr[2, 1]``), slicing (e.g., ``arr[:, 1:5]``), masking (e.g., ``arr[arr > 0]``), fancy indexing (e.g., ``arr[0, [1, 5]]``), and combinations thereof (e.g., ``arr[:, [1, 5]]``).\nHere we'll look at similar means of accessing and modifying values in Pandas ``Series`` and ``DataFrame`` objects.\nIf you have used the NumPy patterns, the corresponding patterns in Pandas will feel very familiar, though there are a few quirks to be aware of.\n\nWe'll start with the simple case of the one-dimensional ``Series`` object, and then move on to the more complicated two-dimesnional ``DataFrame`` object."},{"metadata":{},"cell_type":"markdown","source":"## Data Selection in Series\n\nAs we saw in the previous section, a ``Series`` object acts in many ways like a one-dimensional NumPy array, and in many ways like a standard Python dictionary.\nIf we keep these two overlapping analogies in mind, it will help us to understand the patterns of data indexing and selection in these arrays."},{"metadata":{},"cell_type":"markdown","source":"### Series as dictionary\n\nLike a dictionary, the ``Series`` object provides a mapping from a collection of keys to a collection of values:"},{"metadata":{"trusted":false},"cell_type":"code","source":"import pandas as pd\ndata = pd.Series([0.25, 0.5, 0.75, 1.0],\n                 index=['a', 'b', 'c', 'd'])\ndata","execution_count":1,"outputs":[{"data":{"text/plain":"a    0.25\nb    0.50\nc    0.75\nd    1.00\ndtype: float64"},"execution_count":1,"metadata":{},"output_type":"execute_result"}]},{"metadata":{"trusted":false},"cell_type":"code","source":"data['b']","execution_count":2,"outputs":[{"data":{"text/plain":"0.5"},"execution_count":2,"metadata":{},"output_type":"execute_result"}]},{"metadata":{},"cell_type":"markdown","source":"We can also use dictionary-like Python expressions and methods to examine the keys/indices and values:"},{"metadata":{"trusted":false},"cell_type":"code","source":"'a' in data","execution_count":3,"outputs":[{"data":{"text/plain":"True"},"execution_count":3,"metadata":{},"output_type":"execute_result"}]},{"metadata":{"trusted":false},"cell_type":"code","source":"data.keys()","execution_count":4,"outputs":[{"data":{"text/plain":"Index(['a', 'b', 'c', 'd'], dtype='object')"},"execution_count":4,"metadata":{},"output_type":"execute_result"}]},{"metadata":{"trusted":false},"cell_type":"code","source":"list(data.items())","execution_count":5,"outputs":[{"data":{"text/plain":"[('a', 0.25), ('b', 0.5), ('c', 0.75), ('d', 1.0)]"},"execution_count":5,"metadata":{},"output_type":"execute_result"}]},{"metadata":{},"cell_type":"markdown","source":"``Series`` objects can even be modified with a dictionary-like syntax.\nJust as you can extend a dictionary by assigning to a new key, you can extend a ``Series`` by assigning to a new index value:"},{"metadata":{"trusted":false},"cell_type":"code","source":"data['e'] = 1.25\ndata","execution_count":6,"outputs":[{"data":{"text/plain":"a    0.25\nb    0.50\nc    0.75\nd    1.00\ne    1.25\ndtype: float64"},"execution_count":6,"metadata":{},"output_type":"execute_result"}]},{"metadata":{},"cell_type":"markdown","source":"This easy mutability of the objects is a convenient feature: under the hood, Pandas is making decisions about memory layout and data copying that might need to take place; the user generally does not need to worry about these issues."},{"metadata":{},"cell_type":"markdown","source":"### Series as one-dimensional array"},{"metadata":{},"cell_type":"markdown","source":"A ``Series`` builds on this dictionary-like interface and provides array-style item selection via the same basic mechanisms as NumPy arrays – that is, *slices*, *masking*, and *fancy indexing*.\nExamples of these are as follows:"},{"metadata":{"trusted":false},"cell_type":"code","source":"# slicing by explicit index\ndata['a':'c']","execution_count":7,"outputs":[{"data":{"text/plain":"a    0.25\nb    0.50\nc    0.75\ndtype: float64"},"execution_count":7,"metadata":{},"output_type":"execute_result"}]},{"metadata":{"trusted":false},"cell_type":"code","source":"# slicing by implicit integer index\ndata[0:2]","execution_count":8,"outputs":[{"data":{"text/plain":"a    0.25\nb    0.50\ndtype: float64"},"execution_count":8,"metadata":{},"output_type":"execute_result"}]},{"metadata":{"trusted":false},"cell_type":"code","source":"# masking\ndata[(data > 0.3) & (data < 0.8)]","execution_count":9,"outputs":[{"data":{"text/plain":"b    0.50\nc    0.75\ndtype: float64"},"execution_count":9,"metadata":{},"output_type":"execute_result"}]},{"metadata":{"trusted":false},"cell_type":"code","source":"# fancy indexing\ndata[['a', 'e']]","execution_count":10,"outputs":[{"data":{"text/plain":"a    0.25\ne    1.25\ndtype: float64"},"execution_count":10,"metadata":{},"output_type":"execute_result"}]},{"metadata":{},"cell_type":"markdown","source":"Among these, slicing may be the source of the most confusion.\nNotice that when slicing with an explicit index (i.e., ``data['a':'c']``), the final index is *included* in the slice, while when slicing with an implicit index (i.e., ``data[0:2]``), the final index is *excluded* from the slice."},{"metadata":{},"cell_type":"markdown","source":"### Indexers: loc, iloc, and ix\n\nThese slicing and indexing conventions can be a source of confusion.\nFor example, if your ``Series`` has an explicit integer index, an indexing operation such as ``data[1]`` will use the explicit indices, while a slicing operation like ``data[1:3]`` will use the implicit Python-style index."},{"metadata":{"trusted":false},"cell_type":"code","source":"data = pd.Series(['a', 'b', 'c'], index=[1, 3, 5])\ndata","execution_count":11,"outputs":[{"data":{"text/plain":"1    a\n3    b\n5    c\ndtype: object"},"execution_count":11,"metadata":{},"output_type":"execute_result"}]},{"metadata":{"trusted":false},"cell_type":"code","source":"# explicit index when indexing\ndata[1]","execution_count":12,"outputs":[{"data":{"text/plain":"'a'"},"execution_count":12,"metadata":{},"output_type":"execute_result"}]},{"metadata":{"trusted":false},"cell_type":"code","source":"# implicit index when slicing\ndata[1:3]","execution_count":13,"outputs":[{"data":{"text/plain":"3    b\n5    c\ndtype: object"},"execution_count":13,"metadata":{},"output_type":"execute_result"}]},{"metadata":{},"cell_type":"markdown","source":"Because of this potential confusion in the case of integer indexes, Pandas provides some special *indexer* attributes that explicitly expose certain indexing schemes.\nThese are not functional methods, but attributes that expose a particular slicing interface to the data in the ``Series``.\n\nFirst, the ``loc`` attribute allows indexing and slicing that always references the explicit index:"},{"metadata":{"trusted":false},"cell_type":"code","source":"data.loc[1]","execution_count":14,"outputs":[{"data":{"text/plain":"'a'"},"execution_count":14,"metadata":{},"output_type":"execute_result"}]},{"metadata":{"trusted":false},"cell_type":"code","source":"data.loc[1:3]","execution_count":15,"outputs":[{"data":{"text/plain":"1    a\n3    b\ndtype: object"},"execution_count":15,"metadata":{},"output_type":"execute_result"}]},{"metadata":{},"cell_type":"markdown","source":"The ``iloc`` attribute allows indexing and slicing that always references the implicit Python-style index:"},{"metadata":{"trusted":false},"cell_type":"code","source":"data.iloc[1]","execution_count":16,"outputs":[{"data":{"text/plain":"'b'"},"execution_count":16,"metadata":{},"output_type":"execute_result"}]},{"metadata":{"trusted":false},"cell_type":"code","source":"data.iloc[1:3]","execution_count":17,"outputs":[{"data":{"text/plain":"3    b\n5    c\ndtype: object"},"execution_count":17,"metadata":{},"output_type":"execute_result"}]},{"metadata":{},"cell_type":"markdown","source":"A third indexing attribute, ``ix``, is a hybrid of the two, and for ``Series`` objects is equivalent to standard ``[]``-based indexing.\nThe purpose of the ``ix`` indexer will become more apparent in the context of ``DataFrame`` objects, which we will discuss in a moment.\n\nOne guiding principle of Python code is that \"explicit is better than implicit.\"\nThe explicit nature of ``loc`` and ``iloc`` make them very useful in maintaining clean and readable code; especially in the case of integer indexes, I recommend using these both to make code easier to read and understand, and to prevent subtle bugs due to the mixed indexing/slicing convention."},{"metadata":{},"cell_type":"markdown","source":"## Data Selection in DataFrame\n\nRecall that a ``DataFrame`` acts in many ways like a two-dimensional or structured array, and in other ways like a dictionary of ``Series`` structures sharing the same index.\nThese analogies can be helpful to keep in mind as we explore data selection within this structure."},{"metadata":{},"cell_type":"markdown","source":"### DataFrame as a dictionary\n\nThe first analogy we will consider is the ``DataFrame`` as a dictionary of related ``Series`` objects.\nLet's return to our example of areas and populations of states:"},{"metadata":{"trusted":false},"cell_type":"code","source":"area = pd.Series({'California': 423967, 'Texas': 695662,\n                  'New York': 141297, 'Florida': 170312,\n                  'Illinois': 149995})\npop = pd.Series({'California': 38332521, 'Texas': 26448193,\n                 'New York': 19651127, 'Florida': 19552860,\n                 'Illinois': 12882135})\ndata = pd.DataFrame({'area':area, 'pop':pop})\ndata","execution_count":18,"outputs":[{"data":{"text/html":"<div>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>area</th>\n      <th>pop</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>California</th>\n      <td>423967</td>\n      <td>38332521</td>\n    </tr>\n    <tr>\n      <th>Florida</th>\n      <td>170312</td>\n      <td>19552860</td>\n    </tr>\n    <tr>\n      <th>Illinois</th>\n      <td>149995</td>\n      <td>12882135</td>\n    </tr>\n    <tr>\n      <th>New York</th>\n      <td>141297</td>\n      <td>19651127</td>\n    </tr>\n    <tr>\n      <th>Texas</th>\n      <td>695662</td>\n      <td>26448193</td>\n    </tr>\n  </tbody>\n</table>\n</div>","text/plain":"              area       pop\nCalifornia  423967  38332521\nFlorida     170312  19552860\nIllinois    149995  12882135\nNew York    141297  19651127\nTexas       695662  26448193"},"execution_count":18,"metadata":{},"output_type":"execute_result"}]},{"metadata":{},"cell_type":"markdown","source":"The individual ``Series`` that make up the columns of the ``DataFrame`` can be accessed via dictionary-style indexing of the column name:"},{"metadata":{"trusted":false},"cell_type":"code","source":"data['area']","execution_count":19,"outputs":[{"data":{"text/plain":"California    423967\nFlorida       170312\nIllinois      149995\nNew York      141297\nTexas         695662\nName: area, dtype: int64"},"execution_count":19,"metadata":{},"output_type":"execute_result"}]},{"metadata":{},"cell_type":"markdown","source":"Equivalently, we can use attribute-style access with column names that are strings:"},{"metadata":{"trusted":false},"cell_type":"code","source":"data.area","execution_count":20,"outputs":[{"data":{"text/plain":"California    423967\nFlorida       170312\nIllinois      149995\nNew York      141297\nTexas         695662\nName: area, dtype: int64"},"execution_count":20,"metadata":{},"output_type":"execute_result"}]},{"metadata":{},"cell_type":"markdown","source":"This attribute-style column access actually accesses the exact same object as the dictionary-style access:"},{"metadata":{"trusted":false},"cell_type":"code","source":"data.area is data['area']","execution_count":21,"outputs":[{"data":{"text/plain":"True"},"execution_count":21,"metadata":{},"output_type":"execute_result"}]},{"metadata":{},"cell_type":"markdown","source":"Though this is a useful shorthand, keep in mind that it does not work for all cases!\nFor example, if the column names are not strings, or if the column names conflict with methods of the ``DataFrame``, this attribute-style access is not possible.\nFor example, the ``DataFrame`` has a ``pop()`` method, so ``data.pop`` will point to this rather than the ``\"pop\"`` column:"},{"metadata":{"trusted":false},"cell_type":"code","source":"data.pop is data['pop']","execution_count":22,"outputs":[{"data":{"text/plain":"False"},"execution_count":22,"metadata":{},"output_type":"execute_result"}]},{"metadata":{},"cell_type":"markdown","source":"In particular, you should avoid the temptation to try column assignment via attribute (i.e., use ``data['pop'] = z`` rather than ``data.pop = z``).\n\nLike with the ``Series`` objects discussed earlier, this dictionary-style syntax can also be used to modify the object, in this case adding a new column:"},{"metadata":{"trusted":false},"cell_type":"code","source":"data['density'] = data['pop'] / data['area']\ndata","execution_count":23,"outputs":[{"data":{"text/html":"<div>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>area</th>\n      <th>pop</th>\n      <th>density</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>California</th>\n      <td>423967</td>\n      <td>38332521</td>\n      <td>90.413926</td>\n    </tr>\n    <tr>\n      <th>Florida</th>\n      <td>170312</td>\n      <td>19552860</td>\n      <td>114.806121</td>\n    </tr>\n    <tr>\n      <th>Illinois</th>\n      <td>149995</td>\n      <td>12882135</td>\n      <td>85.883763</td>\n    </tr>\n    <tr>\n      <th>New York</th>\n      <td>141297</td>\n      <td>19651127</td>\n      <td>139.076746</td>\n    </tr>\n    <tr>\n      <th>Texas</th>\n      <td>695662</td>\n      <td>26448193</td>\n      <td>38.018740</td>\n    </tr>\n  </tbody>\n</table>\n</div>","text/plain":"              area       pop     density\nCalifornia  423967  38332521   90.413926\nFlorida     170312  19552860  114.806121\nIllinois    149995  12882135   85.883763\nNew York    141297  19651127  139.076746\nTexas       695662  26448193   38.018740"},"execution_count":23,"metadata":{},"output_type":"execute_result"}]},{"metadata":{},"cell_type":"markdown","source":"This shows a preview of the straightforward syntax of element-by-element arithmetic between ``Series`` objects; we'll dig into this further in [Operating on Data in Pandas](03.03-Operations-in-Pandas.ipynb)."},{"metadata":{},"cell_type":"markdown","source":"### DataFrame as two-dimensional array\n\nAs mentioned previously, we can also view the ``DataFrame`` as an enhanced two-dimensional array.\nWe can examine the raw underlying data array using the ``values`` attribute:"},{"metadata":{"trusted":false},"cell_type":"code","source":"data.values","execution_count":24,"outputs":[{"data":{"text/plain":"array([[  4.23967000e+05,   3.83325210e+07,   9.04139261e+01],\n       [  1.70312000e+05,   1.95528600e+07,   1.14806121e+02],\n       [  1.49995000e+05,   1.28821350e+07,   8.58837628e+01],\n       [  1.41297000e+05,   1.96511270e+07,   1.39076746e+02],\n       [  6.95662000e+05,   2.64481930e+07,   3.80187404e+01]])"},"execution_count":24,"metadata":{},"output_type":"execute_result"}]},{"metadata":{},"cell_type":"markdown","source":"With this picture in mind, many familiar array-like observations can be done on the ``DataFrame`` itself.\nFor example, we can transpose the full ``DataFrame`` to swap rows and columns:"},{"metadata":{"trusted":false},"cell_type":"code","source":"data.T","execution_count":25,"outputs":[{"data":{"text/html":"<div>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>California</th>\n      <th>Florida</th>\n      <th>Illinois</th>\n      <th>New York</th>\n      <th>Texas</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>area</th>\n      <td>4.239670e+05</td>\n      <td>1.703120e+05</td>\n      <td>1.499950e+05</td>\n      <td>1.412970e+05</td>\n      <td>6.956620e+05</td>\n    </tr>\n    <tr>\n      <th>pop</th>\n      <td>3.833252e+07</td>\n      <td>1.955286e+07</td>\n      <td>1.288214e+07</td>\n      <td>1.965113e+07</td>\n      <td>2.644819e+07</td>\n    </tr>\n    <tr>\n      <th>density</th>\n      <td>9.041393e+01</td>\n      <td>1.148061e+02</td>\n      <td>8.588376e+01</td>\n      <td>1.390767e+02</td>\n      <td>3.801874e+01</td>\n    </tr>\n  </tbody>\n</table>\n</div>","text/plain":"           California       Florida      Illinois      New York         Texas\narea     4.239670e+05  1.703120e+05  1.499950e+05  1.412970e+05  6.956620e+05\npop      3.833252e+07  1.955286e+07  1.288214e+07  1.965113e+07  2.644819e+07\ndensity  9.041393e+01  1.148061e+02  8.588376e+01  1.390767e+02  3.801874e+01"},"execution_count":25,"metadata":{},"output_type":"execute_result"}]},{"metadata":{},"cell_type":"markdown","source":"When it comes to indexing of ``DataFrame`` objects, however, it is clear that the dictionary-style indexing of columns precludes our ability to simply treat it as a NumPy array.\nIn particular, passing a single index to an array accesses a row:"},{"metadata":{"trusted":false},"cell_type":"code","source":"data.values[0]","execution_count":26,"outputs":[{"data":{"text/plain":"array([  4.23967000e+05,   3.83325210e+07,   9.04139261e+01])"},"execution_count":26,"metadata":{},"output_type":"execute_result"}]},{"metadata":{},"cell_type":"markdown","source":"and passing a single \"index\" to a ``DataFrame`` accesses a column:"},{"metadata":{"trusted":false},"cell_type":"code","source":"data['area']","execution_count":27,"outputs":[{"data":{"text/plain":"California    423967\nFlorida       170312\nIllinois      149995\nNew York      141297\nTexas         695662\nName: area, dtype: int64"},"execution_count":27,"metadata":{},"output_type":"execute_result"}]},{"metadata":{"collapsed":true},"cell_type":"markdown","source":"Thus for array-style indexing, we need another convention.\nHere Pandas again uses the ``loc``, ``iloc``, and ``ix`` indexers mentioned earlier.\nUsing the ``iloc`` indexer, we can index the underlying array as if it is a simple NumPy array (using the implicit Python-style index), but the ``DataFrame`` index and column labels are maintained in the result:"},{"metadata":{"trusted":false},"cell_type":"code","source":"data.iloc[:3, :2]","execution_count":28,"outputs":[{"data":{"text/html":"<div>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>area</th>\n      <th>pop</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>California</th>\n      <td>423967</td>\n      <td>38332521</td>\n    </tr>\n    <tr>\n      <th>Florida</th>\n      <td>170312</td>\n      <td>19552860</td>\n    </tr>\n    <tr>\n      <th>Illinois</th>\n      <td>149995</td>\n      <td>12882135</td>\n    </tr>\n  </tbody>\n</table>\n</div>","text/plain":"              area       pop\nCalifornia  423967  38332521\nFlorida     170312  19552860\nIllinois    149995  12882135"},"execution_count":28,"metadata":{},"output_type":"execute_result"}]},{"metadata":{},"cell_type":"markdown","source":"Similarly, using the ``loc`` indexer we can index the underlying data in an array-like style but using the explicit index and column names:"},{"metadata":{"trusted":false},"cell_type":"code","source":"data.loc[:'Illinois', :'pop']","execution_count":29,"outputs":[{"data":{"text/html":"<div>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>area</th>\n      <th>pop</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>California</th>\n      <td>423967</td>\n      <td>38332521</td>\n    </tr>\n    <tr>\n      <th>Florida</th>\n      <td>170312</td>\n      <td>19552860</td>\n    </tr>\n    <tr>\n      <th>Illinois</th>\n      <td>149995</td>\n      <td>12882135</td>\n    </tr>\n  </tbody>\n</table>\n</div>","text/plain":"              area       pop\nCalifornia  423967  38332521\nFlorida     170312  19552860\nIllinois    149995  12882135"},"execution_count":29,"metadata":{},"output_type":"execute_result"}]},{"metadata":{"collapsed":true},"cell_type":"markdown","source":"The ``ix`` indexer allows a hybrid of these two approaches:"},{"metadata":{"trusted":false},"cell_type":"code","source":"data.ix[:3, :'pop']","execution_count":30,"outputs":[{"data":{"text/html":"<div>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>area</th>\n      <th>pop</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>California</th>\n      <td>423967</td>\n      <td>38332521</td>\n    </tr>\n    <tr>\n      <th>Florida</th>\n      <td>170312</td>\n      <td>19552860</td>\n    </tr>\n    <tr>\n      <th>Illinois</th>\n      <td>149995</td>\n      <td>12882135</td>\n    </tr>\n  </tbody>\n</table>\n</div>","text/plain":"              area       pop\nCalifornia  423967  38332521\nFlorida     170312  19552860\nIllinois    149995  12882135"},"execution_count":30,"metadata":{},"output_type":"execute_result"}]},{"metadata":{},"cell_type":"markdown","source":"Keep in mind that for integer indices, the ``ix`` indexer is subject to the same potential sources of confusion as discussed for integer-indexed ``Series`` objects.\n\nAny of the familiar NumPy-style data access patterns can be used within these indexers.\nFor example, in the ``loc`` indexer we can combine masking and fancy indexing as in the following:"},{"metadata":{"trusted":false},"cell_type":"code","source":"data.loc[data.density > 100, ['pop', 'density']]","execution_count":31,"outputs":[{"data":{"text/html":"<div>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>pop</th>\n      <th>density</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>Florida</th>\n      <td>19552860</td>\n      <td>114.806121</td>\n    </tr>\n    <tr>\n      <th>New York</th>\n      <td>19651127</td>\n      <td>139.076746</td>\n    </tr>\n  </tbody>\n</table>\n</div>","text/plain":"               pop     density\nFlorida   19552860  114.806121\nNew York  19651127  139.076746"},"execution_count":31,"metadata":{},"output_type":"execute_result"}]},{"metadata":{},"cell_type":"markdown","source":"Any of these indexing conventions may also be used to set or modify values; this is done in the standard way that you might be accustomed to from working with NumPy:"},{"metadata":{"trusted":false},"cell_type":"code","source":"data.iloc[0, 2] = 90\ndata","execution_count":32,"outputs":[{"data":{"text/html":"<div>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>area</th>\n      <th>pop</th>\n      <th>density</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>California</th>\n      <td>423967</td>\n      <td>38332521</td>\n      <td>90.000000</td>\n    </tr>\n    <tr>\n      <th>Florida</th>\n      <td>170312</td>\n      <td>19552860</td>\n      <td>114.806121</td>\n    </tr>\n    <tr>\n      <th>Illinois</th>\n      <td>149995</td>\n      <td>12882135</td>\n      <td>85.883763</td>\n    </tr>\n    <tr>\n      <th>New York</th>\n      <td>141297</td>\n      <td>19651127</td>\n      <td>139.076746</td>\n    </tr>\n    <tr>\n      <th>Texas</th>\n      <td>695662</td>\n      <td>26448193</td>\n      <td>38.018740</td>\n    </tr>\n  </tbody>\n</table>\n</div>","text/plain":"              area       pop     density\nCalifornia  423967  38332521   90.000000\nFlorida     170312  19552860  114.806121\nIllinois    149995  12882135   85.883763\nNew York    141297  19651127  139.076746\nTexas       695662  26448193   38.018740"},"execution_count":32,"metadata":{},"output_type":"execute_result"}]},{"metadata":{},"cell_type":"markdown","source":"To build up your fluency in Pandas data manipulation, I suggest spending some time with a simple ``DataFrame`` and exploring the types of indexing, slicing, masking, and fancy indexing that are allowed by these various indexing approaches."},{"metadata":{},"cell_type":"markdown","source":"### Additional indexing conventions\n\nThere are a couple extra indexing conventions that might seem at odds with the preceding discussion, but nevertheless can be very useful in practice.\nFirst, while *indexing* refers to columns, *slicing* refers to rows:"},{"metadata":{"trusted":false},"cell_type":"code","source":"data['Florida':'Illinois']","execution_count":33,"outputs":[{"data":{"text/html":"<div>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>area</th>\n      <th>pop</th>\n      <th>density</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>Florida</th>\n      <td>170312</td>\n      <td>19552860</td>\n      <td>114.806121</td>\n    </tr>\n    <tr>\n      <th>Illinois</th>\n      <td>149995</td>\n      <td>12882135</td>\n      <td>85.883763</td>\n    </tr>\n  </tbody>\n</table>\n</div>","text/plain":"            area       pop     density\nFlorida   170312  19552860  114.806121\nIllinois  149995  12882135   85.883763"},"execution_count":33,"metadata":{},"output_type":"execute_result"}]},{"metadata":{},"cell_type":"markdown","source":"Such slices can also refer to rows by number rather than by index:"},{"metadata":{"trusted":false},"cell_type":"code","source":"data[1:3]","execution_count":34,"outputs":[{"data":{"text/html":"<div>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>area</th>\n      <th>pop</th>\n      <th>density</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>Florida</th>\n      <td>170312</td>\n      <td>19552860</td>\n      <td>114.806121</td>\n    </tr>\n    <tr>\n      <th>Illinois</th>\n      <td>149995</td>\n      <td>12882135</td>\n      <td>85.883763</td>\n    </tr>\n  </tbody>\n</table>\n</div>","text/plain":"            area       pop     density\nFlorida   170312  19552860  114.806121\nIllinois  149995  12882135   85.883763"},"execution_count":34,"metadata":{},"output_type":"execute_result"}]},{"metadata":{},"cell_type":"markdown","source":"Similarly, direct masking operations are also interpreted row-wise rather than column-wise:"},{"metadata":{"trusted":false},"cell_type":"code","source":"data[data.density > 100]","execution_count":35,"outputs":[{"data":{"text/html":"<div>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>area</th>\n      <th>pop</th>\n      <th>density</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>Florida</th>\n      <td>170312</td>\n      <td>19552860</td>\n      <td>114.806121</td>\n    </tr>\n    <tr>\n      <th>New York</th>\n      <td>141297</td>\n      <td>19651127</td>\n      <td>139.076746</td>\n    </tr>\n  </tbody>\n</table>\n</div>","text/plain":"            area       pop     density\nFlorida   170312  19552860  114.806121\nNew York  141297  19651127  139.076746"},"execution_count":35,"metadata":{},"output_type":"execute_result"}]}],"metadata":{"anaconda-cloud":{},"kernelspec":{"name":"python3","display_name":"Python 3","language":"python"},"language_info":{"name":"python","version":"3.6.11","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat":4,"nbformat_minor":1}