I am on windows 7 64 bit.
I have a csv file 'data.csv'.
I want to import data to a postgresql table 'temp_unicommerce_status' via a python script.
My Script is:
import psycopg2
conn = psycopg2.connect("host='localhost' port='5432' dbname='Ekodev' user='bn_openerp' password='fa05844d'")
cur = conn.cursor()
cur.execute("""truncate table "meta".temp_unicommerce_status;""")
cur.execute("""Copy temp_unicommerce_status from 'C:\Users\n\Desktop\data.csv';""")
conn.commit()
conn.close()
I am getting this error
Traceback (most recent call last):
File "C:\Users\n\Documents\NetBeansProjects\Unicommerce_Status_Update\src\unicommerce_status_update.py", line 5, in <module>
cur.execute("""Copy temp_unicommerce_status from 'C:\\Users\\n\\Desktop\\data.csv';""")
psycopg2.ProgrammingError: must be superuser to COPY to or from a file
HINT: Anyone can COPY to stdout or from stdin. psql's \copy command also works for anyone.
Use the copy_from cursor method
f = open(r'C:\Users\n\Desktop\data.csv', 'r')
cur.copy_from(f, temp_unicommerce_status, sep=',')
f.close()
The file must be passed as an object.
Since you are coping from a csv file it is necessary to specify the separator as the default is a tab character
The way I solved this problem particular to use psychopg2 cursor class function copy_expert (Docs: http://initd.org/psycopg/docs/cursor.html). copy_expert allows you to use STDIN therefore bypassing the need to issue a superuser privilege for the postgres user. Your access to the file then depends on the client (linux/windows/mac) user's access to the file
From Postgres COPY Docs (https://www.postgresql.org/docs/current/static/sql-copy.html):
Do not confuse COPY with the psql instruction \copy. \copy invokes
COPY FROM STDIN or COPY TO STDOUT, and then fetches/stores the data in
a file accessible to the psql client. Thus, file accessibility and
access rights depend on the client rather than the server when \copy
is used.
You can also leave the permissions set strictly for access to the development_user home folder and the App folder.
csv_file_name = '/home/user/some_file.csv'
sql = "COPY table_name FROM STDIN DELIMITER '|' CSV HEADER"
cursor.copy_expert(sql, open(csv_file_name, "r"))
#sample of code that worked for me
import psycopg2 #import the postgres library
#connect to the database
conn = psycopg2.connect(host='localhost',
dbname='database1',
user='postgres',
password='****',
port='****')
#create a cursor object
#cursor object is used to interact with the database
cur = conn.cursor()
#create table with same headers as csv file
cur.execute("CREATE TABLE IF NOT EXISTS test(**** text, **** float, **** float, ****
text)")
#open the csv file using python standard file I/O
#copy file into the table just created
with open('******.csv', 'r') as f:
next(f) # Skip the header row.
#f , <database name>, Comma-Seperated
cur.copy_from(f, '****', sep=',')
#Commit Changes
conn.commit()
#Close connection
conn.close()
f.close()
Here is an extract from relevant PostgreSQL documentation : COPY with a file name instructs the PostgreSQL server to directly read from or write to a file. The file must be accessible to the server and the name must be specified from the viewpoint of the server. When STDIN or STDOUT is specified, data is transmitted via the connection between the client and the server
That's the reason why the copy command to or from a file a restricted to a PostgreSQL superuser : the file must be present on server and is loaded directly by the server process.
You should instead use :
cur.copy_from(r'C:\Users\n\Desktop\data.csv', temp_unicommerce_status)
as suggested by this other answer, because internally it uses COPY from stdin.
You can use d6tstack which makes this simple
import d6tstack
import glob
c = d6tstack.combine_csv.CombinerCSV([r'C:\Users\n\Desktop\data.csv']) # single-file
c = d6tstack.combine_csv.CombinerCSV(glob.glob('*.csv')) # multi-file
c.to_psql_combine('postgresql+psycopg2://psqlusr:psqlpwdpsqlpwd#localhost/psqltest', 'tablename')
It also deals with data schema changes, create/append/replace table and allows you to preprocess data with pandas.
I know this question has been answered, but here are my two cent. I am adding little more description:
You can use cursor.copy_from method :
First you have to create a table with same no of columns as your csv file.
Example:
My csv looks like this:
Name, age , college , id_no , country , state , phone_no
demo_name 22 , bdsu , 1456 , demo_co , demo_da , 9894321_
First create a table:
import psycopg2
from psycopg2 import Error
connection = psycopg2.connect(user = "demo_user",
password = "demo_pass",
host = "127.0.0.1",
port = "5432",
database = "postgres")
cursor = connection.cursor()
create_table_query = '''CREATE TABLE data_set
(Name TEXT NOT NULL ,
age TEXT NOT NULL ,
college TEXT NOT NULL ,
id_no TEXT NOT NULL ,
country TEXT NOT NULL ,
state TEXT NOT NULL ,
phone_no TEXT NOT NULL);'''
cursor.execute(create_table_query)
connection.commit()
Now you can simply use cursor.copy_from where you need three parameters :
first file object , second table_name , third sep type
you can copy now :
f = open(r'final_data.csv', 'r')
cursor.copy_from(f, 'data_set', sep=',')
f.close()
done
I am going to post some of the errors I ran into trying to copy a csv file to a database on a linux based system....
here is an example csv file:
Name Age Height
bob 23 59
tom 56 67
You must install the library psycopg2 (i.e. pip install psycopg2 or sudo apt install python3-psycopg2 )
You must have postgres installed on your system before you can use psycopg2 (sudo apt install postgresql-server postgresql-contrib )
Now you must create a database to store the csv unless you already have postgres setup with a pre-existing database
COPY CSV USING POSTGRES COMMANDS
After installing postgres it creates a default user account which gives you access to postgres commands
To switch to the postgres account issue: sudo -u postgres psql
Acess the prompt by issuing: psql
#command to create a database
create database mytestdb;
#connect to the database to create a table
\connect mytestdb;
#create a table with same csv column names
create table test(name char(50), age char(50), height char(50));
#copy csv file to table
copy mytestdb 'path/to/csv' with csv header;
COPY CSV USING PYTHON
The main issue I ran into with copying the CSV file to a database was I didn't have the database created yet, however this can be done with python still.
import psycopg2 #import the Postgres library
#connect to the database
conn = psycopg2.connect(host='localhost',
dbname='mytestdb',
user='postgres',
password='')
#create a cursor object
#cursor object is used to interact with the database
cur = conn.cursor()
#create table with same headers as csv file
cur.execute('''create table test(name char(50), age char(50), height char(50));''')
#open the csv file using python standard file I/O
#copy file into the table just created
f = open('file.csv','r')
cursor.copy_from(f, 'test', sep=',')
f.close()
Try to do the same as the root user - postgres. If it were linux system, you could change file's permissions or move the file to /tmp. The problem results from missing credentials to read from the filesystem.
Related
import sqlite3
conn = sqlite3.connect("test.db")
cursor = conn.cursor()
It should create the database, but it does not. Any help?
This code will create an sqlite db file called "test.db" in the same directory you are running your script from.
For example, if you have your python file in:
/home/user/python_code/mycode.py
And you run it from:
/home/user/
With:
python python_code/mycode.py # or python3
It will create an "empty" sqlite db file at
/home/user/test.db
If you can't find the test.db file, make sure you pass it the full path of where you want it to be located.
i.e.
conn = sqlite3.connect("/full/path/to/location/you/want/test.db")
I had the same problem, my .db file wasn't appearing because I forgot to add test.db at the end of path, see line 2 below
import sqlite3
databaseFile = "/home/user/test.db" #don't forget the test.db
conn = sqlite3.connect(databaseFile)
cursor = conn.cursor()
I suspect the DB will not be created on disk until you create at least one table in it. Just calling conn.cursor() is not sufficient.
Console sqlite3 utility behaves this way, too.
Is there any way of gaining addressability to the SQLite3 command line interface programmatically from within a script? I would like to write a program that builds a CSV file from a layer attribute table and then imports the CSV file to a database table.
I was thinking that something like this might work:
import sqlite3
conn = sqlite3.connect('qgis.db')
curs = conn.cursor()
commands = '''
DROP Table video IF EXISTS
CREATE TABLE IF NOT EXISTS video(id, name, duration)
.mode csv
.import media.csv video
.mode column
.header on
SELECT * from video
'''
curs.execute(commands)
conn.commit()
conn.close()
I have a Python script where I am trying to merge data from a CSV file into a SQLite spatial database. When I run the UPDATE queries to insert the CSV data into the DB table, they succeed (and I can verify from the python shell that it actually updated the records), but then when I exit Python shell, the database doesn't actually commit the changes. When I exit() and reconnect to the database in a new Python shell, it is back how it was originally, before the UPDATE.
The script is below. The Spatialite DB has all of the land parcels in the county, and information about them (parcel number, address, property values, etc), as well as the vector/polygon data for the parcel's geometry. The CSV file contains three fields PARCEL_NO, OWNER, TAXPAYER. I am trying to, for each parcel number in the database, UPDATE the records for that parcel to include the OWNER and TAXPAYER info from the CSV file in the DB table's OWNER and TAXPAYER fields:
import sqlite3
import csv
dbfile = './ThurstonParcelOwners.sqlite'
ownersfile = './parcel-data/parcels-owners.csv'
parcels = []
conn = sqlite3.connect(dbfile)
c = conn.cursor()
c.execute('''SELECT parcel_no from thurstonparcelowners''')
parcels = [parcel[0] for parcel in c.fetchall()]
owners = []
po = open(ownersfile, 'r')
cw = csv.reader(po, delimiter=',', quotechar='|')
for o in cw:
owners.append(o)
owners = {field[0]:field[1:] for field in owners}
skipped = []
for p in parcels:
try:
owner = owners[p][0]
taxpayer = owners[p][1]
c.execute("""UPDATE thurstonparcelowners SET owner = ? WHERE parcel_no IS ?""", (owner, p))
c.execute("""UPDATE thurstonparcelowners SET taxpayer = ? WHERE parcel_no IS ?""", (taxpayer, p))
except KeyError:
skipped.append(p)
... Again, the UPDATEs are succeeding, and when I run the script with python -i I can verify that it worked afterwards by running SELECT statements on the DB. But when I exit() the Python shell afterwards, the database doesn't retain the data.
Is there something I have to do with sqlite to commit/save the changes before I exit?
In this case, you would have to use conn.commit() to commit the changes to the database.
Writing a script to convert raw data for MySQL import I worked with a temporary textfile so far which I later imported manually using the LOAD DATA INFILE... command.
Now I included the import command into the python script:
db = mysql.connector.connect(user='root', password='root',
host='localhost',
database='myDB')
cursor = db.cursor()
query = """
LOAD DATA INFILE 'temp.txt' INTO TABLE myDB.values
FIELDS TERMINATED BY ',' LINES TERMINATED BY ';';
"""
cursor.execute(query)
cursor.close()
db.commit()
db.close()
This works but temp.txt has to be in the database directory which isn't suitable for my needs.
Next approch is dumping the file and commiting directly:
db = mysql.connector.connect(user='root', password='root',
host='localhost',
database='myDB')
sql = "INSERT INTO values(`timestamp`,`id`,`value`,`status`) VALUES(%s,%s,%s,%s)"
cursor=db.cursor()
for line in lines:
mode, year, julian, time, *values = line.split(",")
del values[5]
date = datetime.strptime(year+julian, "%Y%j").strftime("%Y-%m-%d")
time = datetime.strptime(time.rjust(4, "0"), "%H%M" ).strftime("%H:%M:%S")
timestamp = "%s %s" % (date, time)
for i, value in enumerate(values[:20], 1):
args = (timestamp,str(i+28),value, mode)
cursor.execute(sql,args)
db.commit()
Works as well but takes around four times as long which is too much. (The same for construct was used in the first version to generate temp.txt)
My conclusion is that I need a file and the LOAD DATA INFILE command to be faster. To be free where the textfile is placed the LOCAL option seems useful. But with MySQL Connector (1.1.7) there is the known error:
mysql.connector.errors.ProgrammingError: 1148 (42000): The used command is not allowed with this MySQL version
So far I've seen that using MySQLdb instead of MySQL Connector can be a workaround. Activity on MySQLdb however seems low and Python 3.3 support will probably never come.
Is LOAD DATA LOCAL INFILE the way to go and if so is there a working connector for python 3.3 available?
EDIT: After development the database will run on a server, script on a client.
I may have missed something important, but can't you just specify the full filename in the first chunk of code?
LOAD DATA INFILE '/full/path/to/temp.txt'
Note the path must be a path on the server.
To use LOAD DATA INFILE with every accessible file you have to set the
LOCAL_FILES client flag while creating the connection
import mysql.connector
from mysql.connector.constants import ClientFlag
db = mysql.connector.connect(client_flags=[ClientFlag.LOCAL_FILES], <other arguments>)
I'm trying to modify the two database files used by Google Drive to redirect my sync folder via a script (snapshot.db and sync_conf.db). While I can open the files in certain sqlite browsers (not all) I cant get python to execute a query. I just get the message: sqlite3.DatabaseError: file is encrypted or is not a database
Apparently google is using a Write-Ahead-logging (WAL) configuration on the databases and it can be turned off by running PRAGMA journal_mode=DELETE; (according to sqlite.org) against the database, but I can't figure out how to run that against the database if python can't read it.
heres what I have (I tried executing the PRAGMA command and commiting and then reopening but it didnt work):
import sqlite3
snapShot = 'C:\Documents and Settings\user\Local Settings\Application Data\Google\Drive\snapshot.db'
sync_conf = 'C:\Documents and Settings\user\Local Settings\Application Data\Google\Drive\sync_config.db'
sync_folder_path = 'H:\Google Drive'
conn = sqlite3.connect(snapShot)
cursor = conn.cursor()
#cursor.execute('PRAGMA journal_mode=DELETE;')
#conn.commit()
#conn= sqlite3.connect(snapShot)
#cursor = conn.cursor()
query = "UPDATE local_entry SET filename = '\\?\\" + sync_folder_path +"' WHERE filename ='\\?\C:Users\\admin\Google Drive'"
print query
cursor.execute(query)
problem solved. I just downloaded the latest version of sqlite from http://www.sqlite.org/download.html and overwrote the old .dll in my python27/DLL directory. Works fine now.
What a nusance.
I don't think the journal_mode pragma should keep sqlite3 from being able to open the db at all. Perhaps you're using an excessively old version of the sqlite3 lib? What version of Python are you using, and what version of the sqlite3 library?
import sqlite3
print sqlite3.version