[Improve]:(MongoDB)optimize type inference to avoid unnecessary decimal conversion

[kwonder0926]

Proposed changes

Previously, all numeric types were converted to Decimal to prevent precision loss when syncing mixed-type fields from MongoDB to Doris. This caused unnecessary precision degradation for single-type fields. For example, if the initial table has a small amount of data and an INT-type field with all values below 1000, after conversion to DECIMAL, the numeric precision will be limited to 3 digits. If the data evolves with business growth and the field value later exceeds 1000 (e.g., becomes a 4-digit number), it will lose precision due to the previously inferred limited length.

Problem Summary:

This commit improves the type inference logic:

• Only convert to Decimal if multiple types are present in the sampled data;
• Retain original types (e.g., INT, BIGINT) for fields with consistent data types;

Checklist(Required)

1. Does it affect the original behavior: (No)
2. Has unit tests been added: (Yes)
3. Has document been added or modified: (No Need)
4. Does it need to update dependencies: (No)
5. Are there any changes that cannot be rolled back: (Yes)

Further comments

[JNSimba]

LGTM