[FEA] Handle `binary decimal`s in the main decoder in Parquet reader

**Is your feature request related to a problem? Please describe.**
Currently, the catch-all decoder in Parquet reader is being used to decode the `binary decimal`s, which are not handled by one of our main decoder. Therefore, we should handle `binary decimal`s in our main decoder(s) so that we can move one step closer to purging the catch-all decoder from the codebase.

Related to #18347

**Describe the solution you'd like**
`binary decimal` types being decoded in our main plain decoder instead of in the backup catch-all decoder.

**Describe alternatives you've considered**
N/A

**Additional context**
Here is the parquet file containing binary decimals: [binary_decimal.parquet.zip](https://github.com/user-attachments/files/19477476/binary_decimal.parquet.zip). This file has been generated by the cudf pytest named `test_parquet_reader_binary_decimal`

CC: @GregoryKimball @pmattione-nvidia @nvdbaranec