Order dataframe according to an independent categorical list. Trying to use categories, but having s-CodePudding

The

CodePudding user response：

If I understand your question correctly, what you need to do is turn the column of set names into a categorical column. When doing this you can specify the order you require them to be sorted by. When this is done, sorting by the set names is done in the required order.

Exemplary Input:

mtg_sets_order = [ '15th Anniversary Cards', '2016 Heroes of the Realm', '2017 Gift Pack', '2017 Heroes of the Realm', 'Zendikar Rising Expeditions', 'Zendikar Rising Minigames', 'Zendikar Rising Promos', 'Zendikar Rising Substitute Cards', 'Limited Edition Alpha', 'Unfinity', 'Game Night 2022', 'Limited Edition Beta', 'Unlimited Edition', 'Dominaria United', 'Collectors’ Edition' ] test = ['Limited Edition Alpha', 'Unfinity', 'Game Night 2022', 'Limited Edition Beta', 'Unlimited Edition', 'Dominaria United', 'Collectors’ Edition'] test = pd.DataFrame( columns=['set_name'], data=test )

Code:

# turn set names into categorical column with order as specified in mtg_sets_order test['set_name'] = pd.Categorical(test['set_name'], categories=mtg_sets_order, ordered=True) test.sort_values('set_name')

CodePudding user response：

What you may misunderstand

You may misunderstand the by paramter in of the function sort_value. You may refer to the document. This parameter is which row/column you want sort the whold dataframe according to. And is not used to specify the order used during sorting.

How to sort according to he Categorical

To sort a dataframe according to your custom order which is given by a Categorical, you can refer to this answer.

Basic, if you want to sort a datafrmae d according to it's value a using your custom order order(of type Categorical or List), you can use: d.a = pd.Categorical(df.a, categories=order,ordered=True) to convert the d.a to a pd.Categorical type. As you can see in the document of the Categorical, it's order is unchanged but store the information from order. After this conversion, just call d.sort_values('a', inplace=True)

Code example

The following code can be used to handle the example you provided in the question.

Exemplary Input:

mtg_sets_order = [ '15th Anniversary Cards', '2016 Heroes of the Realm', '2017 Gift Pack', '2017 Heroes of the Realm', 'Zendikar Rising Expeditions', 'Zendikar Rising Minigames', 'Zendikar Rising Promos', 'Zendikar Rising Substitute Cards', 'Limited Edition Alpha', 'Unfinity', 'Game Night 2022', 'Limited Edition Beta', 'Unlimited Edition', 'Collectors’ Edition', 'Dominaria United',] mtg_sets_order = pd.Categorical(mtg_sets_order, categories=mtg_sets_order, ordered=True) test = ['Limited Edition Alpha', 'Unfinity', 'Game Night 2022', 'Limited Edition Beta', 'Unlimited Edition', 'Dominaria United', 'Collectors’ Edition'] test = pd.DataFrame(test)

This is similar to @vogelstein except swapping the last two element of mtg_sets_order. I also cast mtg_sets_order and test to the same type as in your question.

Use the following code to sort according to mtg_sets_order:

test[0] = pd.Categorical(test[0], categories=mtg_sets_order, ordered=True) test.sort_values(by=0, inplace=True)

print(mtg_sets_order) gives:

['15th Anniversary Cards', '2016 Heroes of the Realm', '2017 Gift Pack', '2017 Heroes of the Realm', 'Zendikar Rising Expeditions', ..., 'Game Night 2022', 'Limited Edition Beta', 'Unlimited Edition', 'Collectors’ Edition', 'Dominaria United'] Length: 15 Categories (15, object): ['15th Anniversary Cards' < '2016 Heroes of the Realm' < '2017 Gift Pack' < '2017 Heroes of the Realm' ... 'Limited Edition Beta' < 'Unlimited Edition' < 'Collectors’ Edition' < 'Dominaria United']

print(test) gives:

0 0 Limited Edition Alpha 1 Unfinity 2 Game Night 2022 3 Limited Edition Beta 4 Unlimited Edition 6 Collectors’ Edition 5 Dominaria United

Note that the last two elements are swapped.