From junior to geek the road is sleek
In my case the .csv file had strings as date and time with the following format: 30/07/2017 21:01:17 I find this to be the simplest way to convert this column to float: df[‘dateTime’] = df[‘dateTime’].str.replace(‘/’, ”) df[‘dateTime’] = df[‘dateTime’].str.replace(‘:’, ”) df[‘dateTime’] = df[‘dateTime’].str.replace(‘ ‘, ”) df[‘dateTime’] = df[‘dateTime’].astype(float) So we are removing any signs and … Read more
Hello, I’m back! 😎 Now the traditional method is this: df.loc[df[‘column_name’] == value] For string, and for numeric values this: df.loc[df[‘column_name’] == ‘string’] While it always work with numeric values, for string values sometimes it doesn’t work. It picks up a blank dataframe. I guess it’s something to do with encoding of the source where … Read more
Best way is to use Panda’s to_csv() built-in function df.to_csv(‘filename’, index=False) Don’t forget to pass index=False as a second parameter if you don’t want index row saved as a column (in other words use it if you don’t want duplicated index when opening the file again). By the way, you can also withhold column names … Read more
Adding ‘how=’outer’ to ‘pd.merge()’ will actually do the job! Here’s a very good comprehensive article about Panda’s Join, Merge and Concatenate. And by the way, if you would like to fill NA values when merging, just add ‘.fillna(‘0′)’, in this case it will fill NA values with zeroes.
Results in beautiful Instead of Who’d want to format date like this… 😕
Turns out it’s quite simple if you know the right pandas methods. There’s a simple way to split categorical column into multiple columns with panda’s get_dummies. That is one separate column for each category. Yes, all done automatically. Now we have a bunch of columns with category name in the column names and 0’s and … Read more
Hey guys. Being pretty average at Pandas, yesterday I stumbled upon a formatting challenge. I download some datasheets from the web for machine learning from time to time. This time I got some weird Time & date formatting which might’ve been good for regular use with Excel but unsuitable when it comes to Neural Networks: … Read more