manipulate data of a column - sql

I have a CSV file containing about 30m rows with a column that has a date type. But the problem is with its format. PSQL supports '-' delimiters for timestamp but my dates are using '/'. For example, the date should be '2021-02-01 00:00:00' but my date format is '2021/02/01 00:00:00'. Also, I can not open the CSV file and change it manually due to its large size. I am trying to import my data into a temporary table to replace the '/' with '-' and then inserting them to a new table and I am using the following command(it is not the real table and it is just an example):
CREATE TABLE TMP(
dt VARCHAR
)
CREATE TABLE other_tmp(
dt TIMESTAMP
)
INSERT INTO TMP VALUES('2020/01/02 22:33:11');
INSERT INTO other_tmp(dt)
SELECT dt,
REPLACE(dt, '/', '-')
FROM TMP
I get an error with replace function when I want to run it.
Does anybody know that how can I solve this problem? Or even is it possible to manipulate the column in the original table?


Try this,
In this case, your column "dt" is of type timestamp without time zone but the expression is of type text. So the text should cast to timestamp as below.
CREATE TABLE IF NOT EXISTS TMP
(
dt VARCHAR
);
CREATE TABLE IF NOT EXISTS other_tmp
(
dt TIMESTAMP
);
INSERT INTO TMP
VALUES ('2020/01/02 22:33:11'::TIMESTAMP);
INSERT INTO other_tmp(dt)
SELECT REPLACE(dt, '/', '-')::TIMESTAMP AS dt
FROM TMP;


Your first error is that you have two columns in the SELECT list, but only one column in the INSERT target.
To convert a string to a timestamp, use to_timstamp()
insert into other_tmp(dt)
select to_timestamp(dt, 'yyyy/mm/dd hh24:mi:ss')
from tmp;
Online example

Related

Netezza timestamp data type for column

I'm trying to create a table with one column which is supposed to have a timestamp as the datatype. An example of the value I want to insert into it is this:
2018-12-01T00:00:00Z.
How can I accomplish this? Right now I have it stored as a varchar so when I make joins with this table, in order to convert to timestamp, I need to do this:
to_timestamp(dateColumn, 'YYYY-MM-DDT00:00:00Z)
I'd like to avoid doing this every time I do joins so is there an option maybe when I insert to have this value 2018-12-01T00:00:00Z stored as a Date? It would be fine if the value when inserted into the table appeared as this: 2018-12-01 00:00:00.
EDIT:
Here is my insert statement:
INSERT INTO TEST
SELECT * FROM
EXTERNAL 'C:\...\Desktop\testTable.csv'
USING
(
DELIMITER ','
DATESTYLE 'YMD'
Y2BASE 2000
REMOTESOURCE 'ODBC'
SKIPROWS 1
MAXERRORS 1
LOGDIR 'C:\...\Desktop'
ENCODING 'internal'
DATEDELIM '-'
);
I get the error for the date column: Expected field delimiter or end of record, "2018-12-01" [T]

Bulk insert issue with date from excel into SQL Table

I am trying to bulk insert these two columns from excel into a temp table ##NBP_Table. However, when I do that I get the following error:
'Operand type clash: int is incompatible with date'
Does that mean the date aren't in the format it should be to be inserted into a table?
create table ##NBP_Table
(
Applicable_Date date,
NBP_Value numeric(4,4)
)
insert into ##NBP_Table
values (01/04/2014,1.7107),
(02/04/2014,1.6482),
(03/04/2014,1.686),
(04/04/2014,1.6681)

To get the date insert working, please try this
create table ##NBP_Table
(
Applicable_Date date
NBP_Value numeric(5,4)
)
insert into ##NBP_Table
values ('01/04/2014',1.7107)
The date needs to be in quotation marks
I have also corrected the numeric data type for you

this date in expression is considered as int so it will be performed / operations,
so please use 'before starting date and ' after ending date.
'01-04-2014'
Create table #NBP_Table
(
Applicable_Date date,
NBP_Value numeric(5,4)
)
insert into #NBP_Table
values ('01-04-2014',1.7107),
('02-04-2014',1.6482),
('03-04-2014',1.686),
('04-04-2014',1.6681)

Change Date format during SQL

I am moving data to a new database and need to so some formatting during the move to try and save me time.
The current DB saves the date as YYYYMMDD and the new DB wants it in mm/dd/yyyy. Can I do this during the move or will I have to format after?
Thank you all!

You cold use the format function . So I am not sure how you are getting the data to sql 2014 but once the data is there you could use this command.
This is an example selecting from a table that has a date and altering its format .
use AdventureWorks2012
go
select modifieddate as 'original format', FORMAT ( modifieddate, 'd', 'en-US' ) AS 'New format mm/dd/yy'
from [Sales].[SalesOrderDetail]
Query Result
If your data is not format and its just a string you could use the format command to add the separators .
This code create a table with a date as an INT, the selects the data and formats it as a data time into another table .
CREATE TABLE Test_TimeString (Timeint int)
GO
INSERT INTO Test_TimeString VALUES(04242016)
GO
CREATE TABLE Test_Time (Timedate DATETIME)
GO
INSERT INTO Test_Time
SELECT FORMAT(Timeint,'##/##/####')
FROM Test_TimeString
SELECT * FROM Test_Time

How to insert date values into table

How can I insert into table with different input using / ,with date datatype?
insert into run(id,name,dob)values(&id,'&name',[what should I write here?]);
I'm using oracle 10g.

Since dob is DATE data type, you need to convert the literal to DATE using TO_DATE and the proper format model. The syntax is:
TO_DATE('<date_literal>', '<format_model>')
For example,
SQL> CREATE TABLE t(dob DATE);
Table created.
SQL> INSERT INTO t(dob) VALUES(TO_DATE('17/12/2015', 'DD/MM/YYYY'));
1 row created.
SQL> COMMIT;
Commit complete.
SQL> SELECT * FROM t;
DOB
----------
17/12/2015
A DATE data type contains both date and time elements. If you are not concerned about the time portion, then you could also use the ANSI Date literal which uses a fixed format 'YYYY-MM-DD' and is NLS independent.
For example,
SQL> INSERT INTO t(dob) VALUES(DATE '2015-12-17');
1 row created.

date must be insert with two apostrophes'
As example if the date is 2018/10/20. It can insert from these query
Query -
insert into run(id,name,dob)values(&id,'&name','2018-10-20')

let suppose we create a table Transactions using SQl server management studio
txn_id int,
txn_type_id varchar(200),
Account_id int,
Amount int,
tDate date
);
with date datatype we can insert values in simple format: 'yyyy-mm-dd'
INSERT INTO transactions (txn_id,txn_type_id,Account_id,Amount,tDate)
VALUES (978, 'DBT', 103, 100, '2004-01-22');
Moreover we can have differet time formats like
DATE - format YYYY-MM-DD
DATETIME - format: YYYY-MM-DD HH:MI:SS
SMALLDATETIME - format: YYYY-MM-DD HH:MI:SS

insert into run(id,name,dob)values(&id,'&name',[what should I write
here?]);
insert into run(id,name,dob)values(&id,'&name',TO_DATE('&dob','YYYY-MM-DD'));

You can also use the "timestamp" data type where it just needs "dd-mm-yyyy"
Like:
insert into emp values('12-12-2012');
considering there is just one column in the table...
You can adjust the insertion values according to your table.

I simply wrote an embedded SQL program to write a new record with date fields.
It was by far best and shortest without any errors I was able to reach my requirement.
w_dob = %char(%date(*date));
exec sql insert into Tablename (ID_Number ,
AmendmentNo ,
OverrideDate ,
Operator ,
Text_ID ,
Policy_Company,
Policy_Number ,
Override ,
CREATE_USER )
values ( '801010',
1,
:w_dob,
'MYUSER',
' ',
'01',
'6535435023150',
'1',
'myuser');

To insert the current date you can just use this GETDATE() function.
insert into run(id,name,dob) values(&id,'&name',GETDATE());
you can also use CURRENT_TIMESTAMP() function to insert current date and time.

Alter character field to date

I've a legacy postgres db that has date columns cast as character(50) fields (don't ask). I'd like to alter the table and columns to contain actual dates. Because this worked:
select distinct to_date(date_begin, 'YYYY DD MM') from dates;
I naively thought this might work:
alter table dates alter column date_begin type character
using to_date(date_begin, 'YYYY DD MM');
But it does not. Any clues for the clueless?

This just works as intended by the OP. What we have here is a simple thinko/typo.
Read more in the manual about ALTER TABLE.
Demo:
-- DROP SCHEMA x CASCADE;
CREATE SCHEMA x;
CREATE TABLE x.tbl(date_begin character(50));
INSERT INTO x.tbl VALUES ('2011-11-11 11:11'), (NULL), (''), ('1977');
-- NULL and empty string work too
-- even just YYYY works: '1977' .. is converted to '1977-01-01' automatically
-- empty string produce a possibly surprising result: '0001-01-01 BC'
ALTER TABLE x.tbl ALTER COLUMN date_begin TYPE date USING to_date(date_begin, 'YYYY DD MM');
SELECT * FROM x.tbl;
Hint: You wrote type character instead of type date.

It will take three stages:
1) Alter the table to add the date column.
2) Run an update query which converts each string date into the date field
3) Alter the table to remove the text date column.