BUG: Fix record_prefix ignored when record_path is empty (GH 62205)

[SumitkCodes]

Description

Fixes the bug reported in #62205 where record_prefix was completely ignored when record_path was empty.

The Problem

When calling pd.json_normalize(data, record_prefix="T") with empty record_path, the record_prefix was ignored due to a performance optimization that bypassed the prefix logic.

Before (broken):

import pandas as pd
_s = pd.Series([{"k": f"{i}", "m": "q"} for i in range(5)])
result = pd.json_normalize(_s, record_prefix="T")
# Result: columns were ["k", "m"] (no prefix applied)

Expected behavior:

# Result: columns should be ["T.k", "T.m"] (prefix correctly applied)

The Solution

Modified the early return optimization logic to exclude cases where record_prefix is provided. When record_prefix is provided, the function now properly applies the prefix using the existing nested_to_record function with correct separator handling.

Code Changes

File: pandas/io/json/_normalize.py

The fix ensures that when record_prefix is provided, the function skips the early return optimization and properly applies the prefix, even when record_path is None.

Testing

Added test case test_record_prefix_without_record_path that verifies the bug is fixed. All existing tests continue to pass.

Verification

import pandas as pd
_s = pd.Series([{"k": f"{i}", "m": "q"} for i in range(5)])
result = pd.json_normalize(_s, record_prefix="T")
print(result.columns.tolist())
# Output: ['T.k', 'T.m']

Checklist

• I have checked that this issue has not already been reported
• I have confirmed this bug exists on the latest version of pandas
• I have provided a minimal, reproducible example
• I have confirmed this bug exists on the main branch of pandas
• I have added tests that demonstrate the fix
• All tests pass

Related Issues

Closes #62205