Problem Summary

The current pandas documentation for min_itemsize in HDFStore methods doesn’t clearly explain that it refers to byte length, not character length. This causes confusion when working with multi-byte characters.

Proposed Addition to HDFStore.put() and HDFStore.append() docstrings

Add this clarification to the min_itemsize parameter description in the appropriate methods:

min_itemsize : int, dict, or None, default None
    Minimum size in bytes for string columns when format='table'.
    int - Apply the same minimum size to all string columns, 
    dict - Map column names to their minimum sizes or, 
    None - use the default the sizing
    Important: This specifies the byte length after encoding, not the 
    character count. For multi-byte characters, calculate the required
    size using the encoded byte length.
    See examples below for use.

And adding this to the example section for each docstring:

    Examples:
    - ASCII 'hello' = 5 bytes
    - UTF-8 '香' = 3 bytes (though only 1 character)
    - To find byte length: len(string.encode('utf-8'))

Why This Helps

• Directly addresses the confusion in issue DOC: min_itemsize for HDFStore append for encoded strings #14601
• Provides practical guidance for users working with international text
• Keeps the explanation concise and focused
• Includes actionable examples