As part of the Snowflake roll out we said we would be adding string representations of our IDs to the JSON responses. Today we enabled the string representations and our JSON responses now contain IDs as both a number and the number as a string. Objects which were cached before the string representation was enabled will not contain the _str field until the cache gets updated. To accomodate for this we recommend you write your code to look for the _str data point first, and then fallback to the original ID if the _str doesn't exist.
More information and the developer discussion can be found in the Google Groups thread linked below. Please continue to use that thread for discussions and questions about Snowflake, and check there for code snippets and advice from the Twitter team.
Developer Advocate, Twitter