Home > Enterprise >  Json import to postgresql with python
Json import to postgresql with python

Time:09-09

I'm trying to import multiple json files into my postgres Database with the following code.

import json
import psycopg2


con = psycopg2.connect(database="kw_du", user="postgres", password="f67dd8hzFE5",host="localhost")


cur = con.cursor()
with open(r"C:\Users\diego\OneDrive\Desktop\Workplace\220906\nu_2022_09_01.json", 'r') as data_file:
    data = json.load(data_file)

with con.cursor() as cur:
    cur.executemany('INSERT INTO test_nu(vp_ip, system_activ, vpm_activ) VALUES(%s)', [(json.dumps(d),) for d in data])

con.commit()
con.close()

The json file has the following structure:

[{"vp_id":1,"system_activ":1,"vpm_activ":1},{"vp_id":2,"system_activ":2,"vpm_activ":2}]

When executing the code i get the following error message:

line 13, in cur.executemany('INSERT INTO test_nu (vp_id, system_aktiv) VALUES(%s)', [(json.dumps(d),) for d in data]) psycopg2.errors.SyntaxError: ERROR: INSERT has more target columns than expressions LINE 1: INSERT INTO test_nu (vp_id, system_aktiv) VALUES('{"vp_id": ...

I've seen, that the ERROR: INSERT has more target columns than expressions can have varios reasons, I was not able to solve this problem. Could the ' or {} be a problem? It seems like the json objectr doesn't get parsed VALUES('{"vp_id": ... Appreciate any hint.

CodePudding user response:

You can use json_populate_recordset.

Refer sample query below -

insert into test_nu(vp_id, system_activ, vpm_activ) select * from json_populate_recordset(NULL::test_nu,'[{"vp_id": 1,"system_activ":"eewrw","vpm_activ":"qeqerqr"}]')

Fiddle here.

CodePudding user response:

You can save a number of round-trips to the database with the following:

with expand as (
  select (j->>'vp_id')::int as vp_ip,
         (j->>'system_activ')::int as system_activ,
         (j->>'vpm_activ')::int as vpm_activ
    from jsonb_array_elements((%s)::jsonb) as el(j)
)
insert into test_nu (vp_id, system_activ, vpm_activ)
select vp_ip, system_activ, vpm_activ
  from expand;

Pass json.dumps(data) to this query.

Fiddle Here

  • Related