Pandas DataFrame: Create, Access and Modify

Hello students, welcome to Kwickprep. Your class marks sheet has names down the side and subjects across the top. How can Python hold that whole table and answer questions about it? With a pandas DataFrame. Today we will create a DataFrame, access its values, and change its rows and columns.

Let us start with the new words. Pandas is a Python library for working with tables of data, and we load it with import pandas as p d. A DataFrame is a table with rows and columns, like your marks sheet. The index is the set of row labels, such as roll numbers or names. The column names are the labels across the top. Each single column is a Series, which is a list of values with labels.

The most common way is a dictionary of lists. A dictionary stores key and value pairs inside curly brackets. Each key becomes a column name, and each list becomes that column. P d dot DataFrame builds the table. We did not give row labels, so pandas numbers the rows zero, one and two.

We can also build a DataFrame from Series. Series e has English marks and Series s has Science marks. Each one has marks, with the names as its index. We put them in a dictionary, with Eng for English and Sci for Science as the keys. Pandas matches the rows by their labels, so Riya's marks stay in Riya's row. This time, the row labels are names, not numbers.

A third way is a list of dictionaries, where each dictionary is one row. Notice that Aman's dictionary has no marks. Pandas fills that gap with N a N, which means not a number, a missing value. Because of that missing value, the marks column becomes decimal, so eighty eight shows as eighty eight point zero.

To add a column, write the new column name in square brackets and give it a list. So the Grade column gets A and B. To add a row, use loc with a new row label. The loc tool picks rows by their labels, and we will see it again soon. Neha's row gets sixty four and C.

The rename method changes labels. We give it a dictionary of old name and new name, so Eng becomes Maths. To rename rows, use index instead of columns. The drop method deletes rows or columns. Here it deletes the row Anu and the column Sci. Both methods give back a new DataFrame, so we store it again in df.

Here we set the row labels by giving a list to df dot index. Now there are two ways to reach one value. Loc uses labels, so row Anu and column Sci gives sixty four. Iloc uses integer positions, which start at zero. Row one, column one is also Anu's Science mark. So both lines print sixty four. Find the error: loc of one, Sci gives a key error here, because one is not a row label.

A slice picks a range of rows, written with a colon. Pause and predict. How many rows does loc give from Om to Anu? The answer is two, because a loc slice includes the end label. Iloc from zero to one gives only one row. An iloc slice stops before the end position, just like a list.

Boolean indexing picks rows using a condition. The greater than sign asks if a value is bigger. So df of Eng, greater than seventy, gives True or False for each row. Putting that inside square brackets keeps only the True rows. Raj scored sixty four, so Raj's row is left out.

Attributes are facts about the DataFrame, written after a dot with no brackets. Shape gives rows and columns, so three and two. Size gives the total number of values, six. Columns gives the column labels. Index gives the row labels. D types gives the data type of each column. T swaps rows with columns, which is called the transpose.

To go through a DataFrame one row at a time, use iterrows. Each turn of the loop gives two things. Name is the row label, and row is that row as a Series. So row of Eng plus row of Sci gives each student's total. Om's total is one hundred seventy nine.

Let us revise what we learned today. A DataFrame is a table with labels for rows and columns. You can create it from a dictionary of lists, from Series, or from a list of dictionaries. Add columns with square brackets and rows with loc, then use rename and drop to change them. Loc uses labels, while iloc uses positions. Attributes like shape describe the table, and iterrows goes through each row.