Here's a step-by-step guide on dealing with a Type 2 slowly changing dimension using the MERGE statement in SQL:

Introduction:
A Type 2 slowly changing dimension is used to capture historical changes to dimension attributes while preserving the integrity of the dimension. When new data arrives, we need to handle the insertion of new records while maintaining the historical values. The MERGE statement provides a powerful solution for performing insertions, updates, and handling overlaps in a Type 2 dimension.

Step 1: Design your dimension table:

• Create a dimension table with columns such as the primary key (ID), attributes, effective start date, effective end date, and a flag indicating the current record.

Step 2: Insert new records using MERGE:

• Use the MERGE statement to perform an "upsert" operation, combining insertions and updates into a single statement.
• Specify the dimension table as the target (target) and the new data source as the source (source) in the MERGE INTO clause.
• Define the join condition between the target and source tables based on the primary key.
• When a match is not found (WHEN NOT MATCHED), insert a new record into the dimension table using the values from the source and set the effective start date to the current date, effective end date to a future date (e.g., '9999-12-31'), and the flag indicating the current record to true.

Step 3: Expire existing records:

• Update the existing records in the dimension table that are being updated with the new data.
• Set the effective end date of the current record to the day before the new record's effective start date.
• Set the flag indicating the current record to false.

Step 4: Handle overlaps:

• Identify any overlapping date ranges in the dimension table.
• Update the effective end date of the previous record to the day before the new record's effective start date, ensuring there are no gaps or overlaps.

Here's the SQL script incorporating these steps:

-- Step 2: Insert new records using MERGE
MERGE INTO dimension_table AS target
USING new_data AS source
ON (target.ID = source.ID)
WHEN NOT MATCHED THEN
    INSERT (ID, attribute1, attribute2, effective_start_date, effective_end_date, is_current)
    VALUES (source.ID, source.attribute1, source.attribute2, CURRENT_DATE, '9999-12-31', 1);

-- Step 3: Expire existing records
UPDATE dimension_table
SET effective_end_date = CURRENT_DATE - 1, is_current = 0
WHERE ID IN (
    SELECT ID
    FROM new_data
    INNER JOIN dimension_table ON new_data.ID = dimension_table.ID
    WHERE new_data.attribute1 <> dimension_table.attribute1 OR new_data.attribute2 <> dimension_table.attribute2
);

-- Step 4: Handle overlaps
UPDATE dimension_table AS target
SET effective_end_date = new_record.effective_start_date - 1
FROM (
    SELECT t1.ID, t1.effective_start_date
    FROM dimension_table t1
    INNER JOIN dimension_table t2 ON t1.ID = t2.ID
    WHERE t1.effective_start_date <= t2.effective_start_date
        AND t1.effective_end_date > t2.effective_start_date
        AND t2.is_current = 1
) AS new_record
WHERE target.ID = new_record.ID;

Remember to replace dimension_table with the actual name of your dimension table and adjust the column names and conditions to match your schema.

It's essential to thoroughly test and validate the script to ensure it meets your specific database system and requirements.