إصلاح dtypes والصفوف المكررة

Filling missing values is only half the job. Clean data also needs correct dtypes and no accidental duplicate rows muddying your counts. 🧹

Every column has a dtype, its data type, like int, float, object, or datetime. The dtype controls which operations the column even allows.

A frequent mess: numbers loaded as object strings because one stray symbol crept in. You cannot do math on them until the type is fixed.

Convert a text column to numbers with to_numeric. Pass errors='coerce' to turn anything unparseable into NaN instead of crashing.

When a column is already clean, astype changes its type directly, for example turning whole-number floats back into compact integers.

Dates often arrive as text. Convert them with to_datetime so you can sort, filter by range, and extract parts like month later.

Columns with few repeated text values shrink dramatically as the category dtype, which stores each label once and references it.

Repeated records inflate totals and averages. Find them with duplicated(), which flags each row that has appeared before as True.

Remove repeats with drop_duplicates(). By default it keeps the first occurrence of each row and discards the rest.

Sometimes only certain columns define a duplicate. Pass subset to detect repeats based on a key like an id, ignoring the other columns.

After fixing types and repeats, run info() once more. Confirming dtypes and row counts protects you from cleaning errors slipping downstream.

A price column loaded as object text. Which call safely turns bad entries into NaN?

You now fix dtypes with to_numeric, astype, and to_datetime, and clear repeats with drop_duplicates. Your table is finally analysis-ready.