read_csv และตัวเลือกที่มีประโยชน์

Real analysis starts with a file on disk. The most common one in data science is the humble CSV: plain text, one row per line. 📄

pandas reads a CSV with a single call: read_csv. Hand it a path and you get back a ready-to-use DataFrame.

read_csv returns a full DataFrame: rows, columns, and an automatic integer index. From here every pandas trick is available to you.

Some files split values with a semicolon or tab instead. The sep option tells pandas exactly which character separates your columns.

If the first line is data, not column names, pass header=None so pandas does not steal a real row to use as labels.

Pair header=None with names to supply your own column labels in one clean step. Now every column reads clearly.

The index_col option promotes a column to be the row index, so lookups by that key feel natural later on.

Huge file? The usecols option loads just the columns you list, saving memory and keeping your table focused.

Not sure a file is clean? Use nrows to read a small sample first, inspect it, then load the rest with confidence.

Strange symbols in your text usually mean the wrong encoding. Setting it to utf-8 fixes most accented and non-English characters. ✅

Files use odd tokens for blanks like NA or n/a. The na_values option turns those into real missing values pandas understands.

Your CSV uses semicolons between values. Which option fixes the parse?

You can now load any CSV: choose the separator, set headers and names, pick the index, trim columns, and flag missing values. 🎉