Not all tables available through the Enterprise APIs have been parsed. This is because their confidence score is too low (under 0.5), or because they have a high complexity, such as being a nested table. Even when the API doesn’t parse a table, the API response still informs you that a table is available in the original article data by showing a table identifier. You can extract the raw table data from the wikitext or HTML associated with that article. More info on the table object can be found in the Data Dictionary