PyOrient Client - data_cluster_count()

This method returns the number of records in an array of clusters.

Counting Cluster Records

Using the data_cluster_count() method, you can retrieve a count of the number of records stored on a given cluster. You might find this useful for maintenance tasks or when you need a more granular record count than what the db_count_records() method provides.

Syntax

client.data_cluster_count([<cluster-ids>])
  • <cluster-ids> Defines an array of Cluster ID's.

For more information, see Clusters.

Example

For instance, consider the example of an application to manage smart home devices. In each room you create a physical cluster to store data from environmental sensors. For the web application, you want a count of the number of sensors in an array of rooms.

# Retrieve Record Count by Zone
def zone_count(client, cluster_ids, zone_name):

   # Fetch Record Count
   count = client.data_cluster_count(cluster_ids)

   # Log Findings
   logging.info("Found %s records for %s"
                % (str(count), zone_name))

   # Return
   return count

Where records are organized in clusters by room, for display purposes rooms are grouped together into zones. This function receives the PyOrient Client, an array of Cluster ID's for a given zone and the zone name. Using the Cluster ID's, it fetches a record count for that zone, uses the logging module to report its findings, then passes the count back as a return value.

results matching ""

    No results matching ""