Insert A Dataframe

You can insert a dataframe easily with Relevance AI.


  • Format: the file passed to the insert_csv function must be a valid Pandas dataframe
  • Fields: the dataframe can include as many columns as needed
  • Vector fields: the name of vector fields must end in _vector_
  • Id field: Relevance AI platform identifies unique data entries within a dataset using a field called _id (i.e. every document in the dataset must include an _id field with a unique value per document). There are some arguments to help you take care of this field when using insert_df.

Handling document unique identifier (_id)

  • If the dataset includes a unique identifier per document but the name of the field is not _id, simply pass the name under col_for_id. For instance, in the example below, the field REF-No contains the unique identifier that can be passed as _id.
    Sample data

Using insert_df

First, the Relevance AI SDK package must be installed.

# remove `!` if running the line in a terminal
!pip install -U RelevanceAI[notebook]==2.0.0
from relevanceai import Client

You can sign up/login and find your credentials here:
Once you have signed up, click on the value under `Activation token` and paste it here
client = Client()
ds = client.Dataset('quickstart_insert_df')

Did this page help you?