Подключение к Google Search Console через API, забирает меньше данных

#!/home/user/bigquery-search-console/venv/bin/python
 
import os
import json
import datetime
import requests
import pandas as pd
 
from google.oauth2 import service_account
from googleapiclient.discovery import build
from google.cloud import bigquery
 
#######today_date = '2022-07-15'
today_date_strf = '2022-07-15'
 
########### SET YOUR PARAMETERS HERE ####################################
PROPERTIES = [""]
BQ_DATASET_NAME = ''
BQ_TABLE_NAME = ''
SERVICE_ACCOUNT_FILE = ''
################ END OF PARAMETERS ######################################
 
THIS_FOLDER = os.path.dirname(os.path.abspath(__file__))
file_regal_fortress = os.path.join(THIS_FOLDER, SERVICE_ACCOUNT_FILE)
 
SCOPES = ['https://www.googleapis.com/auth/webmasters.readonly']
credentials = service_account.Credentials.from_service_account_file(
    file_regal_fortress, scopes=SCOPES)
 
service = build(
    'searchconsole',
    'v1',
    credentials=credentials
)
 
 
# Изменил на searchconsole в виду ошибки:
# googleapiclient.errors.UnknownApiNameOrVersion: name: webmasters  version: v3
# service = build(
#     'webmasters',
#     'v3',
#     credentials=credentials
# )
 
def get_sc_df(site_url, start_date, end_date, start_row):
    request = {
        'startDate': start_date,
        'endDate': end_date,
        'dimensions': ["date", "device", "country", "query","page"],
        # uneditable to enforce a nice clean dataframe at the end!
        'rowLimit': 20000,
        'startRow': start_row
    }
 
    response = service.searchanalytics().query(siteUrl=site_url, body=request).execute()
 
    if len(response) > 1:
 
        x = response['rows']
 
        df = pd.DataFrame.from_dict(x)
 
        # split the keys list into columns
        df[["date", "page", "device", "country", "query"]] = pd.DataFrame(df['keys'].values.tolist(), index=df.index)
        df['date'] = pd.to_datetime(df['date'], format='%Y-%m-%d')
 
        # Drop the key columns
        result = df.drop(['keys'], axis=1)
 
        # Add a website identifier
        result['website'] = site_url
 
        # establish a BigQuery client
        client = bigquery.Client.from_service_account_json(file_regal_fortress)
        dataset_id = BQ_DATASET_NAME
        table_name = BQ_TABLE_NAME
        # create a job config
        job_config = bigquery.LoadJobConfig()
        # Set the destination table
        table_ref = client.dataset(dataset_id).table(table_name)
        job_config.destination = table_ref
        job_config.write_disposition = 'WRITE_APPEND'
 
        load_job = client.load_table_from_dataframe(result, table_ref, job_config=job_config)
        load_job.result()
 
        return result
    else:
        print("There are no more results to return.")
 
 
# # Loop through all defined properties, for up to 100,000 rows of data in each
for p in PROPERTIES:
    for x in range(0, 2000000, 20000):
        y = get_sc_df(p, today_date_strf, today_date_strf, x)
        if len(y) < 20000:
            break
        else:
            continue