I am working on a project where I need to create a new table, then import data from a CSV. I've read many similar questions ("extra data after last expected column") and answers on StackOverflow, but I still haven't found the culprit.
CREATE TABLE colleges2014_15 (
unitid integer,
intsnm text,
city text,
stabbr text,
zip_clean char,
control integer,
latitude float,
longitude float,
tutionfee_in float,
tuitionfee_out float,
pctpell float,
inc_pct_lo float,
dep_stat_pct_ind float,
dep_debt_mdn float,
ind_debt_mdn float,
pell_debt_mdn float,
ugds_men float,
ubds_women float,
locale integer,
PRIMARY KEY(unitid)
);
The table is created successfully with the 19 different columns. Then I go try to import the data into the new table.
COPY colleges2014_15(
unitid,
intsnm,
city,
stabbr,
zip_clean,
control,
latitude,
longitude,
tutionfee_in,
tuitionfee_out,
pctpell,
inc_pct_lo,
dep_stat_pct_ind,
dep_debt_mdn,
ind_debt_mdn,
pell_debt_mdn,
ugds_men,
ubds_women,
locale
)
FROM '/Users/compose/Downloads/CollegeScorecard_Raw_Data x/MERGED2014_15_cleaned.csv' CSV HEADER
;
And I get the error message. I've done the following in the CSV:
Made sure it's saved as UTF-8 CSV (working on a Mac)
Already cleaned out all commas in every row
Cleaned out all NULL values
Confirmed that all the data types (integer, float, text, etc.) are correct
I've tried to simply COPY only the first column, unitid; it failed. I've tried importing only the second column (intsnm) and it failed with the same error.
The full error message when trying to COPY over all 19 columns is as follows:
An error occurred when executing the SQL command: COPY
colleges2014_15( unitid, intsnm, city, stabbr, zip_clean,
control, latitude, longitude, tutionfee_in, tuitionfee_out,
pctpell, inc_pct_...
ERROR: extra data after last expected column Where: COPY
colleges2014_15, line 2: "100654,Alabama A & M
University,Normal,AL,35762,35762,1,34.783368,-86.568502,9096,16596,0.7356,0.651..."
1 statement failed.
Execution time: 0.03s
The full error message when trying to copy simply the first column only is:
An error occurred when executing the SQL command: COPY
colleges2014_15( unitid ) FROM
'/Users/compose/Downloads/CollegeScorecard_Raw_Data
x/MERGED2014_15_cleaned.csv' CSV HEADER
ERROR: extra data after last expected column Where: COPY
colleges2014_15, line 2: "100654,Alabama A & M
University,Normal,AL,35762,35762,1,34.783368,-86.568502,9096,16596,0.7356,0.651..."
1 statement failed.
Execution time: 0.01s
Hugely appreciate any help.
It took me a while to figure out what was wrong searching on the error so have posted my problem to help others. My issue was inexperience with pgAdmin as
pgAdmin requires the table to be created WITH columns before the data is imported. I had expected that the headers would be used from the .csv file, most other packages I have used worked this way.
If you are working with a GIS system using PostGIS there is an easy solution. I am using QGIS 3.4, with Postgres and PostGIS installed.
In QGIS
Select Database menu option
Select DBManager
On left - choose location for table
Select Import Layer/File
On the next window select the following
Input - choose file
Table - enter table name
OK
Your CSV has a ZIP column which your table and COPY statement omit.
Related
The maximum size of limited character types (e.g. varchar(n)) in Postgres is 10485760.
description on max length of postgresql's varchar
Please download the file for testing and extract it in /tmp/2019q4, we only use pre.txt to import data with.
sample data
Enter you psql and create a database:
postgres=# create database edgar;
postgres=# \c edgar;
Create table according to the webpage:
fields in pre table definations
edgar=# create table pre(
id serial ,
adsh varchar(20),
report numeric(6,0),
line numeric(6,0),
stmt varchar(2),
inpth boolean,
rfile char(1),
tag varchar(256),
version varchar(20),
plabel varchar(512),
negating boolean
);
CREATE TABLE
Try to import data:
edgar=# \copy pre(adsh,report,line,stmt,inpth,rfile,tag,version,plabel,negating) from '/tmp/2019q4/pre.txt' with delimiter E'\t' csv header;
We analyse the error info:
ERROR: value too long for type character varying(512)
CONTEXT: COPY pre, line 1005798, column plabel: "LIABILITIES AND STOCKHOLDERS EQUITY 0
0001493152-19-017173 2 11 BS 0 H LiabilitiesAndStockholdersEqu..."
Time: 1481.566 ms (00:01.482)
1.What size i set in the field is just 512 ,more less than 10485760.
2.the content in line 1005798 is not same as in error info:
0001654954-19-012748 6 20 EQ 0 H ReclassificationAdjustmentRelatingToAvailableforsaleSecuritiesNetOfTaxEffect 0001654954-19-012748 Reclassification adjustment relating to available-for-sale securities, net of tax effect" 0
Now i drop the previous table ,convert the plabel field as text,re-create it:
edgar=# drop table pre;
DROP TABLE
Time: 22.763 ms
edgar=# create table pre(
id serial ,
adsh varchar(20),
report numeric(6,0),
line numeric(6,0),
stmt varchar(2),
inpth boolean,
rfile char(1),
tag varchar(256),
version varchar(20),
plabel text,
negating boolean
);
CREATE TABLE
Time: 81.895 ms
Import the same data with same copy command:
edgar=# \copy pre(adsh,report,line,stmt,inpth,rfile,tag,version,plabel,negating) from '/tmp/2019q4/pre.txt' with delimiter E'\t' csv header;
COPY 275079
Time: 2964.898 ms (00:02.965)
edgar=#
No error info in psql console,let me check the raw data '/tmp/2019q4/pre.txt' ,which it contain 1043000 lines.
wc -l /tmp/2019q4/pre.txt
1043000 /tmp/2019q4/pre.txt
There are 1043000 lines,how much lines imported then?
edgar=# select count(*) from pre;
count
--------
275079
(1 row)
Why so less data imported without error info ?
The sample data you provided is obviously not the data you are really loading. It does still show the same error, but of course the line numbers and markers are different.
That file occasionally has double quote marks where there should be single quote marks (apostrophes). Because you are using CSV mode, these stray double quotes will start multi-line strings, which span all the way until the next stray double quote mark. That is why you have fewer rows of data than lines of input, because some of the data values are giant multiline strings.
Since your data clearly isn't CSV, you probably shouldn't be using \copy in CSV format. It loads fine in text format as long as you specify "header", although that option didn't become available in text format until v15. For versions before that, you could manually remove the header line, or use PROGRAM to skip the header like FROM PROGRAM 'tail +2 /tmp/pre.txt' Alternatively, you could keep using CSV format, but choose a different quote character, one that never shows up in your data such as with (delimiter E'\t', format csv, header, quote E'\b')
I'm very confused about this error I'm getting in my Query Tool in PgAdmin. I've been working on this for days, and cannot find a solution to fixing this error when attempting to upload this csv file to my Postgres table.
ERROR: invalid input syntax for type numeric: "2021-02-14"
CONTEXT: COPY CardData, line 2, column sold_price: "2021-02-14"
SQL state: 22P02
Here is my code in the Query Tool that I am running
CREATE TABLE Public."CardData"(Title text, Sold_Price decimal, Bids int, Sold_Date date, Card_Link text, Image_Link text)
select * from Public."CardData"
COPY Public."CardData" FROM 'W:\Python_Projects\cardscrapper_project\ebay_api\card_data_test.csv' DELIMITER ',' CSV HEADER ;
Here is a sample from the first row of my csv file.
Title,Sold_Date,Sold_Price,Bids,Card_Link,Image_Link
2018 Contenders Optic Sam Darnold #103 Red Rookie #/99 PSA 8 NM-MT AUTO 10,2021-02-14,104.5,26,https://www.ebay.com/itm/2018-Contenders-Optic-Sam-Darnold-103-Red-Rookie-99-PSA-8-NM-MT-AUTO-10/143935698791?hash=item21833c7767%3Ag%3AjewAAOSwNb9gGEvi&LH_Auction=1,https://i.ebayimg.com/thumbs/images/g/jewAAOSwNb9gGEvi/s-l225.jpg
The "Sold_Date" column is in the correct datetime format that is easy for Postgres to understand, but the error is calling on the "Sold-Price" column?
I'm very confused. Any help is greatly appreciated.
Notice that the columns are not in the same order in the csv file and in the table.
You would have to specify the proper column order
COPY Public."CardData" (Title,Sold_Date,Sold_Price,Bids,Card_Link,Image_Link)
FROM 'W:\Python_Projects\cardscrapper_project\ebay_api\card_data_test.csv'
DELIMITER ',' CSV HEADER ;
You have created the table with sold_price as the second column, so the COPY command will expect a price/number to be the second column in your CSV file. Your CSV file, however has sold_date as the second column, which will lead to the data type mismatch error that you see.
Either you can re-define your CREATE TABLE statement with the sold_date as second column and sold_price as 4th column, or you can specify the column parsing order in your COPY statement as COPY public."CardData" (<column order>)
Another option is to open up the CSV file in Excel and re-order the columns and do a Save As...
I'm trying to use the copy function to create a table in Redshift. I've setup this particular field that keeps failing in my schema as a standard timestamp because I don't know why it would be anything otherwise. But when I run this statement:
copy sample_table
from 's3://aws-bucket/data_push_2018-10-05.txt'
credentials 'aws_access_key_id=XXXXXXXXXXXXXXXXXXXX;aws_secret_access_key=XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX/XXX'
dateformat 'auto'
ignoreheader 1;
It keeps returning this error: Invalid timestamp format or value [YYYY-MM-DD HH24:MI:SS]
raw_field_value: "2018-08-29 15:04:52"
raw_line: 12039752|311525|"67daf211abbe11e8b0010a28385dd2bc"|98953|"2018-08-20"|"2018-11-30"|"active"|"risk"|||||||"sample"|15750|0|"2018-08-29 15:04:52"|"2018-08-29 16:05:01"
There is a very similar table in our database (that I did not make) which has the aforementioned error value as timestamp and values for that field identical to 2018-08-29 15:04:52 so what's happening when I run it that's causing the issue?
Your copy command seems OK, and seems like you are missing FORMAT as CSV QUOTE AS '"' AND DELIMITER AS '|' parameters and It should work.
I'm here using some sample data and command to prove my case, to make it simple, I did made the table simple and covered all your data points though.
create table sample_table(
salesid integer not null,
category varchar(100),
created_at timestamp,
update_at timestamp );
Here goes your sample data test_file.csv,
12039752|"67daf211abbe11e8b0010a28385dd2bc"|"2018-08-29 11:04:52"|"2018-08-29 14:05:01"
12039754|"67daf211abbe11e8b0010a2838cccddbc"|"2018-08-29 15:04:52"|"2018-08-29 16:05:01"
12039755|"67daf211abbe11e8b0010a28385ff2bc"|"2018-08-29 12:04:52"|"2018-08-29 13:05:01"
12039756|"67daf211abbe11e8b0010a28385bb2bc |"2018-08-29 10:04:52"|"2018-08-29 15:05:01"
Here goes your copy command,
COPY sample_table FROM 's3://path/to/csv/test_file.csv' CREDENTIALS 'aws_access_key_id=XXXXXXXXXXX;aws_secret_access_key=XXXXXXXXX' FORMAT as CSV QUOTE AS '"' DELIMITER AS '|';
It will returns,
INFO: Load into table 'sample_table' completed, 4 record(s) loaded successfully.
COPY
Though this command works fine, but if there are more issues with your data you could try MAXERROR option as well.
Hope it answers your question.
I'm trying to import a CSV file to my PostgreSQL but I get this error
ERROR: invalid input syntax for integer: "id;date;time;latitude;longitude"
CONTEXT: COPY test, line 1, column id: "id;date;time;latitude;longitude"
my csv file is simple
id;date;time;latitude;longitude
12980;2015-10-22;14:13:44.1430000;59,86411203;17,64274849
The table is created with the following code:
CREATE TABLE kordinater.test
(
id integer NOT NULL,
date date,
"time" time without time zone,
latitude real,
longitude real
)
WITH (
OIDS = FALSE
)
TABLESPACE pg_default;
ALTER TABLE kordinater.test
OWNER to postgres;
You can use Import/Export option for this task.
Right click on your table
Select "Import/Export" option & Click
Provide proper option
Click Ok button
You should try this it must work
COPY kordinater.test(id,date,time,latitude,longitude)
FROM 'C:\tmp\yourfile.csv' DELIMITER ',' CSV HEADER;
Your csv header must be separated by comma NOT WITH semi-colon or try to change id column type to bigint
to know more
I believe the quickest way to overcome this issue is to create an intermediary temporary table, so that you can import your data and cast the coordinates as you please.
Create a similar temporary table with the problematic columns as text:
CREATE TEMPORARY TABLE tmp
(
id integer,
date date,
time time without time zone,
latitude text,
longitude text
);
And import your file using COPY:
COPY tmp FROM '/path/to/file.csv' DELIMITER ';' CSV HEADER;
Once you have your data in the tmp table, you can cast the coordinates and insert them into the test table with this command:
INSERT INTO test (id, date, time, latitude, longitude)
SELECT id, date, time, replace(latitude,',','.')::numeric, replace(longitude,',','.')::numeric from tmp;
One more thing:
Since you're working with geographic coordinates, I sincerely recommend you to take a look at PostGIS. It is quite easy to install and makes your life much easier when you start your first calculations with geospatial data.
I am using Postgres 9.5.3(On Ubuntu 16.04) and I have a table with some timestamptz fields
...
datetime_received timestamptz NULL,
datetime_manufactured timestamptz NULL,
...
I used the following SQL command to generate CSV file:
COPY (select * from tmp_table limit 100000) TO '/tmp/aa.csv' DELIMITER ';' CSV HEADER;
and used:
COPY tmp_table FROM '/tmp/aa.csv' DELIMITER ';' CSV ENCODING 'UTF-8';
to import into the table.
The example of rows in the CSV file:
CM0030;;INV_AVAILABLE;2016-07-30 14:50:42.141+07;;2016-08-06 00:00:000+07;FAHCM00001;;123;;;;;1.000000;1.000000;;;;;;;;80000.000000;;;2016-07-30 14:59:08.959+07;2016-07-30 14:59:08.959+07;2016-07-30 14:59:08.959+07;2016-07-30 14:59:08.959+07;
But I encounter the following error when running the second command:
ERROR: invalid input syntax for type timestamp with time zone: "datetime_received"
CONTEXT: COPY inventory_item, line 1, column datetime_received: "datetime_received"
My database's timezone is:
show timezone;
TimeZone
-----------
localtime(GMT+7)
(1 row)
Is there any missing step or wrong configuration?
Any suggestions are appreciated!
The error you're seeing means that Postgres is trying (and failing) to convert the string 'datetime_received' to a timestamp value.
This is happening because COPY is trying to insert the header row into your table. You need to include a HEADER clause on the COPY FROM command, just like you did for the COPY TO.
More generally, when using COPY to move data around, you should make sure that the TO and FROM commands are using exactly the same options. Specifying ENCODING for one command and not the other can lead to errors, or silently corrupt data, if your client encoding is not UTF8.