问题
In an existing PostgreSQL table, I would like to UPDATE
several existing columns with values from a dictionary lookup (see dict below). Somewhat like described in this nice blog post. However, I can't figure out how to do that with a Python dictionary. Here comes the terrible pseudo-code:
d = {10:'chair', 11:'table', 12:'lamp',
20:'english ivy', 21:'peace lily', 22:'spider plant'}
curs.execute("""
UPDATE my_table t
SET furniture = %(t.furniture)s,
SET plant = %(t.plant)s""",
d)
The original table would look somewhat like this:
gid | furniture | plant
-----------------------
0 | 10 | 21
1 | 11 | 20
...
After the operation it should look like this:
gid | furniture | plant
-----------------------------
0 | chair | peace lily
1 | table | english ivy
...
Is this possible or will I have to loop through the table?
回答1:
try this:
rows = (
{'gid': 10, 'furniture': 10, 'plant': 10},
{'gid': 20, 'furniture': 20, 'plant': 20}
)
cur.executemany(
'''
UPDATE myTable
SET
furniture = %(furniture)s,
plant = %(plant)s
WHERE
gid = %(gid)s
''',
rows
)
回答2:
The approach of catver works. However, I found that creating a temporary table proved to be more efficient.
import psycopg2
from psycopg2.extensions import AsIs
rows = zip(d.keys(), d.values())
curs.execute("""
CREATE TEMP TABLE codelist(DKEY INTEGER, DVALUE TEXT)
ON COMMIT DROP""")
curs.executemany("""
INSERT INTO codelist (DKEY, DVALUE)
VALUES(%s, %s)""",
rows)
for i in [(AsIs('furniture'), AsIs('furniture')), (AsIs('plant'), AsIs('plant'))]:
curs.execute("""
UPDATE my_table
SET %s = codelist.DVALUE
FROM codelist
WHERE codelist.DKEY = my_table.%s;
""", i)
NB: This example may not quite work because I am replacing INTEGER
with TEXT
values. This may throw the error ERROR: operator does not exist: integer = character varying
.
In that case, this answer might help.
来源:https://stackoverflow.com/questions/30162121/bulk-update-postgresql-from-python-dict