I want week number from different columns of Year, Month & Day in same table of Hive QL.
IF you can give logic on SQL, that's also fine.
Concatenate the date, month and year into a proper date format and apply weekofyear().
Select weekofyear(cast(concat(year,"-",month,"-",date) as date)) from tablename.
Please note that I have used cast to convert the concatenated string into date.Howver, you might need to use different method based on your date format. Please refer to the below answer on handling string conversion to date formats.
Hive cast string to date dd-MM-yyyy
Related
Sorry - this may be a basic question, but I have been banging my head against this for a week.
I have a database field with the format "dd/mm/yyyy hh:mm:ss" called UpdateTime and referencing max(AuditHistory.ActionedDateTime) in the database.
I am trying to identify the Week / ISO Week from the date part of this field only using the dataset in ReportBuilder3.
I am trying to achieve an integer entry in a column called "WeekNo" giving me the week of the year that a transaction was made so I can use this for grouping results by year | by week number on a report for senior management.
I have tried many combinations of:
,DATEPART(WEEK,DAY(max(AuditHistory.ActionedDateTime)) AS WeekNo and
,DATEPART(WEEK,MONTH(max(AuditHistory.ActionedDateTime)) AS WeekNo.
If I use a static date, e.g. , DATEPART(WEEK,DAY('1900-01-20')) AS WeekNo, it returns perfectly as "4" but I cannot for the life of me get the datepart format correct to identify the week from the format of the field.
I believe my issue is getting SQL to accept that the field is "dd/mm/yyyy hh:mm:ss" and work out the week from the date element.
Before I go mad - I thought I'd ask if there is a quick way to achieve this.
The DATEPART function expects a date / datetime / datetime2 value. You are passing in an integer representing the day or month number.
Assuming you're storing your dates correctly, you just need to pass in the date value directly:
DATEPART(WEEK, Max(AuditHistory.ActionedDateTime)) As WeekNo
I read other similar question where the answer is to use FORMAT(getdate(),'yyy-MM') or something similar. However the problem for me in using anything like this, is that it changes the date type to a varchar or char. I need it to stay as datetype but just want Year and Month.. I tried the following..-> FORMAT(a.completeddate,'yyy-MM') which works to change to year and month but the date is no longer a datetype or date format. So when I try to do the following -> select #FirstCompletion = (A.completeddate) i get this error..Conversion failed when converting date and/or time from character string. Basically I need to convert the date column to year and month as date format so I can then pass values to variables using select #FirstCompletion = (A.completeddate) and set #secondMonth = DATEADD(month, 2, #FirstCompletion) which are Datetype variables.. Would appreciate any help I can get.. Thanks..
I am writing the query to show the data entries for a specific year. Date is stored in dd/mm/yyyy hh:mm:ss.(Date TIMESTAMP - e.g. 12/2/2014 0:00:00).
I am trying to display the two columns(name, orderdate) filtered by a specific year(year from orderdate). The requirement is to enter the specific year(2010 or 2020 etc) not the entire date. I tried using date_format() and regexp_replace() with WHERE but nothing helped.
Can someone help me?
If your are storing the date -- incorrectly -- as a string, then you can use string functions to do what you want:
where orderdate like '__/__/2010%'
However, you should really put your effort into storing the date using a correct format -- YYYY-MM-DD for strings at least.
I have a problem with converting a varchar2 fields into a date format.
I got 2 columns with the datatyp varchar2, one is called qtime the other is called ztime. Both fields contain strings in this format (f.e. 152015 -> would be a timestamp 15:20:15).
For reporting reasons I need to convert this fields into a date format, afterwards I want to substract (qtime-ztime) the fields an convert them into the format [hh] (f.e. after the operation 01:20:00 would be -> 01). Is it possible to to this within Oracle SQL 12c? The biggest problem for me right now is that I don't get those Strings converted into a date format.
select TO_DATE(qtime,'MM/DD/YYYY hh24:mi:ss') just gives me
ORA-01861:"literal does not match format string"
select TO_DATE(qtime,'hh24mmss') gives me a wrong Date
01.03.2018
select TO_TIMESTAMP(qtime,'hh24mmss') gives me a wrong Date
01.03.2018 BUT the correct time with f.e. 15:20:15,0000000
Thank you in advance, any help is appreciated
Note: I only have reading rights on the database Oracle 12c, so I need to to this within Statements
"The Database contains another column with the correct date for each time"
The missing piece of the puzzle! Concatenate the two columns to get something which can be converted to an Oracle DATE:
select to_date(qdate||qtime, 'yyyymmddhh24miss') as qdatetime
, to_date(zdate||ztime, 'yyyymmddhh24miss') as zdatetime
from your_table
Once you have done that you can perform arithmetic of the dates e.g.
select id
, zdatetime - qdatetime as time_diff
from ( select id
, to_date(qdate||qtime, 'yyyymmddhh24miss') as qdatetime
, to_date(zdate||ztime, 'yyyymmddhh24miss') as zdatetime
from your_table
)
If you want the number of hours in the difference you can include this expression in the projection of the outer query:
, extract( hour from (zdatetime - qdatetime) day to second) as hrs_ela
First off, if you are trying to convert a varchar2 into a date without specifying neither day nor month, it will default to the first day of the current month:
If you specify a date value without a date, then the default date is the first day of the current month.
You can read up more here
Also in 2nd and 3rd your example, you are using 'hh24mmss' for specifying hour, minute and second components, but do note that correct format for minutes is 'mi' and not 'mm', which is used for months.
So the solution is to concatenate both date and time component when creating the date as the other answer suggested, tho I would recommend using a single date field as it can store the information you need.
I am using sql for database connectivity. I am filtering records in database on date basis using sql query which consists of function convert(varchar(100),column name,106). I am comparing it with date having dd/MMM/yyyy format. As a result I get records of next all the dates.
Example: If I try to filter the table by 01/June/2017 then inspite of having records of that date, I get records of next all the dates. I am not able to get records of 01/June/2017.
I have used greater than or equal to expression.
It is better to convert it to this yyyymmdd number format and do the comparison as below:
select convert(int,convert(varchar(10),[your column],112))
20170529
Don't convert your dates into any format (=strings) when doing comparison. Always use the date (or datetime etc) format. That way you will be comparing the actual values, not string or numeric comparison. Here the problem is that format 106 is dd mon yyyy so it will never match your date because it's a string dd/MMM/yyyy but you will get any dates starting with 02 anyhow, for example 02/Jan/1990, because in alphabetical order that will come after your given date.