how to format date column in kdb - date

I am newbie to KDB. I have a KDB table which I am querying as:
select[100] from table_name
now this table has got some date columns which have dates stored in this format
yyyy.mm.dd
I wish to query that table and retrieve the date fields in specific format (like mm/dd/yyyy). If this would've been any other RDBMS table this is what i would have done:
select to_date(date_field,'mm/dd/yyyy') from table_name
I need kdb equivalent of above. I've tried my best to go through the kdb docs but unable to find any function / example / syntax to do that.
Thanks in advance!

As Anton said KDB doesn't have an inbuilt way to specify the date format. However you can extract the components of the date individually and rearrange as you wish.
For the example table t with date column:
q)t
date
----------
2008.02.04
2015.01.02
q)update o:{"0"^"/"sv'flip -2 -2 4$'string`mm`dd`year$\:x}date from t
date o
-----------------------
2008.02.04 "02/04/2008"
2015.01.02 "01/02/2015"
From right to left inside the function: we extract the month,day and year components with `mm`dd`year$:x before stringing the result. We then pad the month and day components with a null character (-2 -2 4$') before each and add the "/" formatting ("/"sv'flip). Finally the leading nulls are filled with "0" ("0"^).

Check out this GitHub library for datetime formatting. It supports the excel way of formatting date and time. Though it might not be the right fit for formatting a very large number of objects (but if distinct dates are very less then a keyed table and lj can be used for lookup).
q).dtf.format["mm/dd/yyyy"; 2016.09.23]
"09/23/2016"
q).dtf.format["dd mmmm yyyy"; 2016.09.03] // another example
"03 September 2016"

I don't think KDB has built-in date formatting features.
The most reliable way is to format date by yourself.
For example
t: ([]date: 10?.z.d);
update dateFormatted: {x: "." vs x; x[1],"/",x[2],"/",x[0]} each string date from t
gives
date dateFormatted
------------------------
2012.07.21 "07/21/2012"
2001.05.11 "05/11/2001"
2008.04.25 "04/25/2008"
....
Or, more efficient way to do the same formatting is
update dateFormatted: "/"sv/:("."vs/:string date)[;1 2 0] from t

now qdate is available for datetime parsing and conversion

Related

In snowflake , how to convert one date format to another format. From YYYYMMDD to YYYY-MON-DD

I have table ABC in which I have column Z of datatype Date. The format of the data is YYYYMMDD. Now I am looking to convert the above format to YYYY-MON-DD format. Can someone help?
You can use to_char
TO_CHAR(Z,'YYYY-MON-DD')
Depending on what the purpose of the reformatting is, you can either explicitly cast it to a VARCHAR/CHAR and define the format, or you can change your display format to however you'd like to see all dates:
ALTER SESSION SET DATE_OUTPUT_FORMAT = 'YYYY-MON-DD';
It's important to understand that if the data is in a DATE field, then it is stored as a date, and the format of the date is dependent on your viewing preferences, not how it is stored.
Since the value of the date field is stored as a number, you have to convert it to date.
ALTER SESSION SET DATE_OUTPUT_FORMAT = 'YYYY-MON-DD';
select to_date(to_char( z ), 'YYYYMMDD');
(adding this answer to summarize and resolve the question - since the clues and answers are scattered through comments)
The question stated that column Z is of type DATE, but it really seems to be a NUMBER.
Then before parsing a number like 20201017 to a date, first you need to transform it to a STRING.
Once the original number is parsed to a date, it can be represented as a new string formatted as desired.
WITH data AS (
SELECT 20201017 AS z
)
SELECT TO_CHAR(TO_DATE(TO_CHAR(z), 'YYYYMMDD'), 'YYYY-MON-DD')
FROM data;
# 2020-Oct-17

To filter sales based on max date value minus 6 months

I'm trying to filter the data based on the date filter. The date column in my table is in date9. format(30JUN2017).
I want to filter the date column by subtracting 6 months from the existing date, which will be 31DEC2017.
e.g: date<31DEC2017
Can anyhow suggest on how to do this, I have tried using intx function and other options as well.
Thanks for your help.
Best Regards,
AJ
When manipulating dates in SAS using their formatted values you should enclose them in ""d, i.e. in your example it should be where date<"31DEC2017"d. If your date is stored in date9 format as a string, you can make it into a date like so: where input(date,date9.)<"31DEC2017"d
Now your question seems to differ slightly from its title. According to the latter, you want to filter on max_date - 6 months, whatever that max date might be in your dataset. This could be done so:
proc sql;
select *
from t
where intnx('month',date,6)<(select max(date) from t)
;
quit;
If however, for some strange reason, by "max date" you mean the current date, the above query simply becomes this:
proc sql;
select *
from t
where intnx('month',date,6)<date()
;
quit;

Is ISO8601 the best date-format for PostgreSQL jsonb when i want to filter by the date?

I'm new to PostgreSQL and I have the following question:
I have a table with just an id-column and a data-column, which uses the jsonb-type. Inside the jsonb-object I have a datetime field. I read in various posts, that I should use the ISO-8601 dateformat to store in the DB.
I want to filter my table by date like this:
SELECT * FROM table WHERE data->'date' > '2016-01-01T00:00'
Is this really the best date-format for this purpose?
Thanks in advance :)
IMHO Your query should produce
ERROR: operator does not exist: jsonb > timestamp with time zone
If I get it right. In case you change -> to ->> it should get a text value instead of jsonb field (which is also not comparable to timestamp).
It should be smth like
SELECT * FROM table WHERE (data->>'date')::timestamptz > '2016-01-01T00:00' to work
The big advantage of that format is that string order corresponds to date order, so a comparison like the one you quote in your question would actually work as intended.
A second advantage is that a timestamp in that format can easily be converted to a PostgreSQL timestamp with time zone value, because the type input function understands this format.
I hope you are not dealing with dates “before Christ”, because it wouldn't work so easily with those.

Postgres timestamp to date

I am building a map in CartoDB which uses Postgres. I'm simply trying to display my dates as: 10-16-2014 but, haven't been able to because Postgres includes an unneeded timestamp in every date column.
Should I alter the column to remove the timestamp or, is it simply a matter of a (correct) SELECT query? I can SELECT records from a date range no problem with:
SELECT * FROM mytable
WHERE myTableDate >= '2014-01-01' AND myTableDate < '2014-12-31'
However, my dates appear in my CartoDB maps as: 2014-10-16T00:00:00Z and I'm just trying to get the popups on my maps to read: 10-16-2014.
Any help would be appreciated - Thank you!
You are confusing storage with display.
Store a timestamp or date, depending on whethether you need time or not.
If you want formatted output, ask the database for formatted output with to_char, e.g.
SELECT col1, col2, to_char(col3, 'DD-MM-YY'), ... FROM ...;
See the PostgreSQL manual.
There is no way to set a user-specified date output format. Dates are always output in ISO format. If PostgreSQL let you specify other formats without changing the SQL query text it'd really confuse client drivers and applications that expect the date format the protocol specifies and get something entirely different.
You have two basic options.
1 Change the column from a timestamp to a date column.
2 Cast to date in your SQL query (i.e. mytimestamp::date works).
In general if this is a presentation issue, I don't usually think that is a good reason to muck around with the database structure. That's better handled by client-side processing or casting in an SQL query. On the other hand if the issue is a semantic one, then you may want to revisit your database structure.

Convert a Date field into Week Number in IBM DB2

I have a Date field (CHAR Datatype) which has values in this format: YYYYMMDD.
For example 20140729.
I want to convert it into a Weeknumber in format YYYYWKNO
For example, the result would be 201432.
How this can be done in IBM DB2?
Luckily for you, DB2 has some pretty good date formatting functions. Although the links for the documentation are for DB2 for Linux/Unix/Windows, it will also work on DB2 for z/OS.
You can use TIMESTAMP_FORMAT() to convert your CHAR field to an actual date, which you can then use VARCHAR_FORMAT() to format it in the way you wish:
SELECT
VARCHAR_FORMAT(
TIMESTAMP_FORMAT(
'20140801'
,'YYYYMMDD'
)
,'YYYYWW'
)
FROM SYSIBM.SYSDUMMY1
There are two different formats for "week", one is WW, which will give the week based on a week beginning with January 1 and ending January 7, and IW, which will give the ISO Week.. Please see the documentation page for VARCHAR_FORMAT for the other formats available.
The following gets week 31 instead of 32. But overall I think it is a good solution for this problem:
SELECT TO_DATE('20140801', 'YYYYMMDD') AS MYDATE
, YEAR(TO_DATE('20140801', 'YYYYMMDD')) * 100 + WEEK(TO_DATE('20140801', 'YYYYMMDD')) AS YYYYWKNO
FROM SYSIBM.SYSDUMMY1