Check if the Synapse pipeline has any transformations or mappings that are altering the schema. Sometimes during data ingestion, pipelines might drop or rename columns if they are not properly configured.
If the schema inference does not detect all the columns or there is a mismatch between data types, it may omit some columns. You can try manually specifying the schema or using schema mapping in your dataflow.
Data masking or security policies in Dataverse could also prevent some columns from being extracted under certain conditions. Ensure that there are no row-level security (RLS) or column-level security (CLS) policies in place that restrict certain columns from being visible during extraction.