CODEWITHSUNDEEP
performanceorientdbpyorient
laetitil
laetitil

Reputation: 11

How to implement efficiently the pattern "create if not exist" in Orientdb?

I want to insert in Orientdb unique objects but in order to avoid duplicate object I make a query and then I create my object if it does not exist. I have billions of objects to insert and this is very taken a long time. How to avoid duplicate object in insertion and have good performance?

Here is a sample of my code (I'm using pyorient by the way) :

# creation object Address src
        query_ip_src = client.query("select @rid from `Address` where address_value  = '" + log_value[2] + "' parallel")
        if len(query_ip_src) == 0:
            ip_src = Address()
            ip_src.address_value = log_value[2]

            ip_src_record = client.record_create(clusters[b'address'], ip_src.to_dict())
            ip_src_rid = str(ip_src_record._rid)
        else:
            ip_src_rid = "#" + str(query_ip_src[0].rid.get())

Upvotes: 0

Views: 330

Answers (1)

Luigi Dell'Aquila
Luigi Dell'Aquila

Reputation: 2814

There is an UPDATE UPSERT SQL statement for this, eg.

UPDATE Address SET address_value = ?, otherField = ? UPSERT WHERE address_value = ?

Just make sure you have a unique index on the relevant unique fields (address_value in this case), this will prevent data duplication

Upvotes: 1

Related Questions