df.select_dtypes(include=none, exclude=none)
- Select all numeric types, use np.number or number
- Select all strings, use object type
- 选择datetimes, 用datetime or datetime64
- 选择 timedeltas, 用 hour delta
- Select pandas categorical type, use 'category' ??
ex:
df.select_dtypes (include = ['object'])
Here, the selected columns are all string type columns in the dataframe
~ In Python, it means the opposite, a bit similar to sql! =
Select some rows in the DataFrame, you can use ~ to get False from True
dropna: delete the row record with null value
dropna (subset = ['xx', 'xxx']): find free worth records in the xx and xxx columns, and delete
Python-utils
- It is a toolkit that integrates many python functions and commonly used or reusable classes.
- When we create the code, we can also put the function or class that can be reused in utils and call it together with the main code
df.empty: returns True, if the dataframe is empty, otherwise returns False
Counter (): In Python 3, it is used to map the value of the required number
ex:
c = Counter () # Create a new empty counter
c = Counter ('abcasdf') # A counter generated by an iterative object
c = Counter ({'red': 4, 'yello': 2}) # A map generated counter
c = Counter (cats = 2, dogs = 5) # counter generated by keyword parameter
generates counter, although it is not useful herefrom collections import Counter
c = Counter(‘abcasd’)
c
Counter({‘a’: 2, ‘c’: 1, ‘b’: 1, ‘s’: 1, ‘d’: 1})c2 = Counter©
c2
Counter({‘a’: 2, ‘c’: 1, ‘b’: 1, ‘s’: 1, ‘d’: 1})