pivot vs pivot_table in pandas
`pivot` reshapes unique values only, while `pivot_table` can aggregate duplicates.
- pivot needs unique index-column pairs
- pivot_table accepts aggfunc
- pivot_table is safer for real data
Quick recall
374 cards
read_csv performance tips in pandas
Limit parsed columns, set dtypes intentionally, and stream large files in chunks when needed.
- usecols avoids unused data
- dtype reduces inference cost
- chunksize helps large files
Series vs DataFrame in pandas
A Series is one labeled column, while a DataFrame is a 2D labeled table of columns.
- Series is 1D
- DataFrame is tabular
- Both align by index
What causes SettingWithCopyWarning in pandas?
It happens when pandas is unsure whether you are writing to a view or a detached copy.
- Chained indexing is the common trigger
- Use loc for assignment
- Make copies explicit when needed
What does explode do in pandas?
`explode` turns each item in a list-like cell into its own row while repeating the other column values.
- List-like values become multiple rows
- Other columns repeat
- Useful before groupby or joins
What does vectorization mean in pandas?
Vectorization means using array-style operations on whole columns instead of Python loops over rows.
- Operate on columns, not row-by-row
- Usually faster than iterrows
- Leverages NumPy underneath
When should you use categorical dtype in pandas?
Use categorical dtype for repeated low-cardinality labels to save memory and make category semantics explicit.
- Good for repeated labels
- Smaller than object strings
- Categories can be ordered
Working with datetime columns in pandas
Convert early with `to_datetime`, then use the `.dt` accessor for filtering, grouping, and feature extraction.
- Normalize text to datetime first
- dt exposes year month weekday and more
- Time zones matter
Advantages of the Django ORM
The ORM improves productivity with composable queries, model validation, and database abstraction.
- Readable query API
- Built-in escaping reduces SQL injection risk
- Easy relation traversal
Class-based views vs function-based views in Django
Function-based views are explicit and simple, while class-based views reduce repetition through reusable behavior and inheritance.
- FBVs are straightforward
- CBVs are reusable
- Choose based on clarity not fashion
Django caching basics
Cache expensive work at the right layer: low-level data, per-view output, or template fragments.
- Trade freshness for speed
- Use stable cache keys
- Invalidate intentionally
Explain Django MVT architecture
Django splits concerns into models, views, and templates, with the framework routing requests between them.
- Model holds data rules
- View handles request logic
- Template renders response