Mongo also stores the same column/attribute names with every single row. 50000 inserts? 50000 (usually identical) sets of attribute names stored (compared to once for SQL). And to my knowledge they are stored uncompressed. For a Big Data database, it doesn't seem very good at efficiently storing big data.
This could be a problem, for sure. There are some tools on top of MongoDB that tries to reduce it, for example mongoengine will allow you to define a "compressed name" to store in the DB, meaning that you'll see 'timestamp' on your codebase, but that will be stored as 't'.
Comments
Mongo also stores the same column/attribute names with every single row. 50000 inserts? 50000 (usually identical) sets of attribute names stored (compared to once for SQL). And to my knowledge they are stored uncompressed. For a Big Data database, it doesn't seem very good at efficiently storing big data.
This could be a problem, for sure. There are some tools on top of MongoDB that tries to reduce it, for example mongoengine will allow you to define a "compressed name" to store in the DB, meaning that you'll see 'timestamp' on your codebase, but that will be stored as 't'.