使用包含空格的列名称或使用包含空格的列名称的drop方法查询Pandas DataFrame

Question

I am looking to use pandas to drop rows based on the column name (contains a space) and the cell value. I have tried various ways to achieve this (drop and query methods) but it seems I'm failing due to the space in the name. Is there a way to query the data using the name that has a space in it or do I need to clean all spaces first?

data in form of a csv file

Date,"price","Sale Item"
2012-06-11,1600.20,item1
2012-06-12,1610.02,item2
2012-06-13,1618.07,item3
2012-06-14,1624.40,item4
2012-06-15,1626.15,item5
2012-06-16,1626.15,item6
2012-06-17,1626.15,item7

Attempt Examples

df.drop(['Sale Item'] != 'Item1')
df.drop('Sale Item' != 'Item1')
df.drop("'Sale Item'] != 'Item1'")

df.query('Sale Item' != 'Item1')
df.query(['Sale Item'] != 'Item1')
df.query("'Sale Item'] != 'Item1'")

Error received in most cases

ImportError: 'numexpr' not found. Cannot use engine='numexpr' for query/eval if 'numexpr' is not installed

Answer 1

If I understood correctly your issue, maybe you can just apply a filter like:

df = df[df['Sale Item'] != 'item1']

which returns:

         Date    price Sale Item
1  2012-06-12  1610.02     item2
2  2012-06-13  1618.07     item3
3  2012-06-14  1624.40     item4
4  2012-06-15  1626.15     item5
5  2012-06-16  1626.15     item6
6  2012-06-17  1626.15     item7

Answer 2

As you can see from the documentation -

DataFrame.drop(labels, axis=0, level=None, inplace=False, errors='raise')

Return new object with labels in requested axis removed

DataFrame.drop() takes the index of the rows to drop, not the condition. Hence you would most probably need something like -

df.drop(df.ix[df['Sale Item'] != 'item1'].index)

Please note, this drops the rows that meet the condition, so the result would be the rows that don't meet the condition, if you want the opposite you can use ~ operator before your condition to negate it.

But this seems a bit too much, it would be easier to just use Boolean indexing to get the rows you want (as indicated in the other answer) .

---

Demo -

In [20]: df
Out[20]:
         Date    price Sale Item
0  2012-06-11  1600.20     item1
1  2012-06-12  1610.02     item2
2  2012-06-13  1618.07     item3
3  2012-06-14  1624.40     item4
4  2012-06-15  1626.15     item5
5  2012-06-16  1626.15     item6
6  2012-06-17  1626.15     item7

In [21]: df.drop(df.ix[df['Sale Item'] != 'item1'].index)
Out[21]:
         Date   price Sale Item
0  2012-06-11  1600.2     item1