I85PREP
⌘ K
RBSE SS Informatics Practices study materials
RBSE SS Informatics Practices study resources

RBSE SS Informatics Practices: Pandas: CSV Import/Export, Cleaning, Joining and Descriptive Statistics

RBSE SS Informatics Practices · Work through your exam, one topic at a time.

All guide sections

Pandas: CSV Import/Export, Cleaning, Joining and Descriptive Statistics

What this domain tests

This domain tests the full small-data workflow: read a CSV into a DataFrame, inspect and clean it, combine two tables, summarise it with statistics, and write results back to disk.

Import and export

  • pd.read_csv('data.csv') treats the first row as column headers and uses a comma separator by default. Useful arguments are sep, header=None, names=[...], index_col, usecols, nrows, na_values and dtype.
  • df.to_csv('out.csv', index=False) prevents the row index from being written as an extra unnamed column.

Cleaning and joining

  • Inspect with head(), tail(), shape, info(), isnull().sum() and duplicated().sum().
  • Clean with dropna(how='any'/'all', subset=[...]), fillna(value), drop_duplicates(subset=..., keep='first'), rename(columns={...}), astype() and str.strip().
  • pd.merge(left, right, on='key', how='inner'/'left'/'right'/'outer') combines tables on a shared column; pd.concat([df1, df2], axis=0) stacks rows and axis=1 places columns side by side.

Descriptive statistics

df.describe() reports count, mean, standard deviation, minimum, quartiles and maximum for numeric columns only; add include='all' to include text columns. Series methods such as mean(), median(), mode(), value_counts() and corr() give finer summaries.

Worked example

A students table has Roll and Name; a marks table has Roll and Marks. pd.merge(students, marks, on='Roll', how='left') keeps every student, filling NaN for anyone with no marks. With how='inner' those students would disappear. Reasoning: the join key matches rows; the join type decides which unmatched rows survive.

Traps that cost marks

Forgetting index=False in to_csv; assuming describe() covers text columns; confusing merge with concat; expecting dropna() to remove only fully empty rows; filling missing values without checking whether the mean is a sensible substitute.