Replacing values on a dataframe can sometimes be very tricky. Bulk replacement in a large dataset could be difficult and slow.<\/p>\n\n\n\n
Yet, Pandas is flexible enough to do it better.<\/p>\n\n\n\n
You may have to postpone the idea of moving to a distributed computing system. At least not to solve the value replacement problems.<\/p>\n\n\n\n
Related: Is Your Python For-loop Slow? Use NumPy Instead<\/a><\/em><\/strong><\/p>\n\n\n\n The popular approach to replacing a value is locating the cell (or the column) and assigning it a new value.<\/p>\n\n\n\n Here’s how people do it (and how I did it when starting data analysis.)<\/p>\n\n\n\n The above is a dataset of top Instagram influencers. You can download the original dataset from Kaggle<\/a>. Let’s suppose we’ve noticed that the dataset has “N\/A” in place of some missing values.<\/p>\n\n\n\n People will do this.<\/p>\n\n\n\nReplacing column values: The old approach.<\/h2>\n\n\n\n
Title<\/th> Category<\/th> Followers<\/th> Engagement<\/th><\/tr><\/thead> Cristiano Ronaldo<\/td> Sports with a ball<\/td> 400100000.0<\/td> 7800000.0<\/td><\/tr> Kylie ????<\/td> Fashion|Modeling|Beauty<\/td> N\/A<\/td> 6200000.0<\/td><\/tr> Leo Messi<\/td> Sports with a ball|Family<\/td> 306300000.0<\/td> N\/A<\/td><\/tr> Kendall<\/td> Modeling|Fashion<\/td> N\/A<\/td> 3400000.0<\/td><\/tr> Selena Gomez<\/td> Music|Lifestyle<\/td> 295800000.0<\/td> 2700000.0<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n