Question about documentation for Synaptic Connections Database

[sprucey8]

For the code in the synaptic connections database (under C3 segmentation layers in the Released data section of the H01 release website), I am wondering how the different connectivity data is split up between the files? For example, is the data in one file corresponding to neurons in a certain cortical layer? Or is split up by neuron type? Thank you for any suggestions!

[timblakely]

Heya!

The database was exported directly from an internal BigQuery table. In order to allow for parallelization during export, BigQuery splits the table at arbitrary-but-roughly-size-balanced locations and then encodes them in Avro format. There is no guarantee that a specific synapse or neuron ID is in any one of the given shards. Ideally you can ingest the table back into your own BigQuery dataset and use that to query it (or into a local SQL database if you have one available).