# Get data from pipedrive API and insert it into Snowflake

I am a new python programmer. I have created the code to put API data from pipedrive to Snowflake database. Here are the steps in my code.

1. Delete the csv file if it exists.
2. Make an API call put all the paginated data in a list.
3. Create a csv file from the list.
4. Truncate the table in Snowflake.
5. Remove data from Snowflake stage table
6. Put the data in Snowflake stage table.
7. Copy data from stage table to a normal table.

I would love to get some feedback on it as I will create more scripts based on this code.

Here is my code.

import requests
from module import usr, pwd, acct, db, schem, api_token
import snowflake.connector
import datetime
import time
from datetime import datetime
import csv
import os
import contextlib

end_point = 'persons'
limit = 500
start = 0
start_time = time.time()
csvfile = r'C:/Users/User1/PycharmProjects/Pipedrive/persons.csv'
def snowflake_connect():
mydb = snowflake.connector.connect(
user=usr,
account=acct,
database=db,
schema=schem,
)
cursor = mydb.cursor()
return cursor

def snowflake_truncate(cursor):
print("Truncating table PERSONS_NEW: {}".format(datetime.now().strftime("%Y-%m-%d %H:%M:%S")))
cursor.execute('TRUNCATE TABLE PERSONS_NEW')
print("PERSONS_NEW truncated: {}".format(datetime.now().strftime("%Y-%m-%d %H:%M:%S")))
return cursor

def snowflake_insert(cursor):
cursor.execute("remove @persons_test pattern='.*.csv.gz'")
for c in cursor:
print(c, datetime.now().strftime("%Y-%m-%d %H:%M:%S"))
cursor.execute('put file://{} @persons_test auto_compress=true'.format(csvfile))
for c in cursor:
print(c, datetime.now().strftime("%Y-%m-%d %H:%M:%S"))
cursor.execute("""COPY INTO MARKETING.PIPEDRIVE_MASTER.persons_new FROM @persons_test/persons.csv.gz file_format=(TYPE=csv field_delimiter=',' skip_header=0 FIELD_OPTIONALLY_ENCLOSED_BY = '"') on_error = 'abort_statement'""")
for c in cursor:
print(c, datetime.now().strftime("%Y-%m-%d %H:%M:%S"))

def get_persons(start):
url = 'https://company.pipedrive.com/v1/{}?user_id=0&start={}&limit={}&api_token={}'.format(end_point, start, limit, api_token)
response = requests.request("GET", url).json()
url = 'https://company.pipedrive.com/v1/{}?user_id=0&start={}&limit={}&api_token={}'.format(end_point, start, limit, api_token)
response = requests.request("GET", url).json()
start = start + 500

for data in response['data']:
id = data['id']
activities_count = data['activities_count']
else:
closed_deals_count = data['closed_deals_count']
company_id = data['company_id']
done_activities_count = data['done_activities_count']
followers_count = data['followers_count']
label = data['label']
last_activity_date = data['last_activity_date']
last_activity_id = data['last_activity_id']
last_incoming_mail_time = data['last_incoming_mail_time']
last_name = data['last_name']
last_outgoing_mail_time = data['last_outgoing_mail_time']
lost_deals_count = data['lost_deals_count']
name = data['name']
next_activity_date = data['next_activity_date']
next_activity_id = data['next_activity_id']
next_activity_time = data['next_activity_time']
notes_count = data['notes_count']
open_deals_count = data['open_deals_count']
if data['org_id'] == None:
org_id = None
else:
org_id = data['org_id']['value']
org_name = data['org_name']

fieldnames = [id, activities_count, add_time, cc_email, closed_deals_count, company_id, done_activities_count, followers_count, label, last_activity_date, last_activity_id, last_incoming_mail_time,
last_name, last_outgoing_mail_time, lost_deals_count, name, next_activity_date, next_activity_id, next_activity_time, notes_count, open_deals_count, org_id, org_name]
write_csv(fieldnames)

def delete_existing_csv():
with contextlib.suppress(FileNotFoundError):
os.remove(csvfile)

def write_csv(fieldnames):
with open(csvfile, "a", encoding="utf-8", newline='') as fp:
wr = csv.writer(fp, delimiter=',')
wr.writerow(fieldnames)

if __name__ == "__main__":
delete_existing_csv()
print("Creating CSV file: {}".format(datetime.now().strftime("%Y-%m-%d %H:%M:%S")))
get_persons(start)
print("CSV file succesfully created: {}".format(datetime.now().strftime("%Y-%m-%d %H:%M:%S")))
cursor = snowflake_connect()
snowflake_truncate(cursor)
snowflake_insert(cursor)
cursor.close()
end_time = time.time()
elapsed_time = round(end_time - start_time, 2)
print("Job sucessfully completed in: {} seconds".format(elapsed_time))


The script looks pretty good to me.

The biggest issue is that if you plan to execute this code multiple times, it's not really efficient. I think you should check for existing persons in Snowflake, and then only add the persons that are not yet in there. I'm not sure how Snowflake works so I can't help you with that, but my approach would be to create a Person class, create a __hash__ and an __eq__ method, and put all existing persons in a set. Then for each person you read from Pipedrive, check if it is in the Persons set, and if it's not, add it to Snowflake. This would prevent costly truncate operations plus a lot of insert operations, in case you have a lot of persons in your Pipedrive.

Next, add some comments. The code should be self-explanatory so you could incorporate the steps you mentioned above as comments.

Finally, regarding global variables (PEP dictates they be capitalized):

• Rename limit to PIPEDRIVE_PAGINATION_LIMIT
• Remove the start variable and initialize get_persons(start=0)
• Put start_time under if __name__ == "__main__"
• Rename csv_file to PIPEDRIVE_PERSONS_CSV

For the commenting and global variables stuff, try linting your code, for example using pylint. It'll give you hints on what you can improve about your code.

• Thanks for the feedback. I will change my variables according to PEP. I know multiple truncates and inserts are not efficient in general. But in this case I have only few thousands rows of data. This data can change. So instead of keeping track of updates for more than 30 columns (here I just put a small set of fields) for a small data set I decided to truncate and load. It takes average 11 seconds to do that with Snowflake stage table. – jmf Jan 30 at 15:30