summing time column in mysql - mysql

I have a column in one of my tables, which is TIME format (00:00:00). I am trying to sum the entire column and display it as same (00:00:00).
I have tried using the following but it is not giving me anywhere near the correct answer.It's giving me 22.12:44:00 and manual calcaulation tells me it should be close to 212:something:something
SELECT SEC_TO_TIME( SUM( TIME_TO_SEC( vluchttijd ) ) ) AS totaltime FROM tbl_vluchtgegevens
Any recommendations?

You can try like this:-
SELECT SEC_TO_TIME(SUM(SECOND(vluchttijd ))) AS totaltime FROM tbl_vluchtgegevens;
or try this(althoug this is not a good approach):
SELECT concat(floor(SUM( TIME_TO_SEC( `vluchttijd ` ))/3600),":",floor(SUM( TIME_TO_SEC( `vluchttijd ` ))/60)%60,":",SUM( TIME_TO_SEC( `vluchttijd ` ))%60) AS total_time
FROM tbl_vluchtgegevens;
Edit:-
Try this:-
select cast(sum(datediff(second,0,dt))/3600 as varchar(12)) + ':' +
right('0' + cast(sum(datediff(second,0,dt))/60%60 as varchar(2)),2) +
':' + right('0' + cast(sum(datediff(second,0,dt))%60 as varchar(2)),2)
from TestTable
Working SQL Fidlle

In MySQL, the TIME type is rather limited in range. Moreover many time function do not accept values greater that 23:59:59, making it really usable only to represent the time of the day.
Given your needs, your best bet is probably to write a custom function that will mimic SEC_TO_TIME but allowing much greater range:
CREATE FUNCTION SEC_TO_BIGTIME(sec INT)
RETURNS CHAR(10) DETERMINISTIC
BEGIN
SET #h = sec DIV 3600;
SET #m = sec DIV 60 MOD 60;
SET #s = sec MOD 60;
RETURN CONCAT(
LPAD(#h, 4, '0'),
':',
LPAD(#m, 2, '0'),
':',
LPAD(#s, 2, '0')
);
END;
And here is how to use it:
create table tbl (dt time);
insert tbl values
('09:00:00'), ('01:00:00'), ('07:50:15'), ('12:00:00'),
('08:30:00'), ('00:45:00'), ('12:10:30');
select SEC_TO_BIGTIME(sum(time_to_sec(dt))) from tbl;
Producing:
+--------------------------------------+
| SEC_TO_BIGTIME(SUM(TIME_TO_SEC(DT))) |
+--------------------------------------+
| 0051:15:45 |
+--------------------------------------+
See http://sqlfiddle.com/#!8/aaab8/1
Please note the result is a CHAR(10) in order to overcome TIMEtype limitations. Depending how you plan to use that result, that means that you may have to convert from that string to the appropriate type in your host language.

This worked for me:
SELECT SEC_TO_TIME(SUM(TIME_TO_SEC(vluchttijd))) AS totaltime FROM tbl_vluchtgegevens;

Related

MySql search integer range from number with dash

I have table in that I have one field with dash value. Like...
I need to search this with between condition.
For example if I have one value 25 then I need to search the records which include the value 25 like 20-31. In above image there are 6 records which include 25 value. So it should return 6 records.
Please help me in this query ? What would be the query for that ?
You can use MySQL's substring_index() function to easily get the data before and after the dash:
select substring_index(yourcolumn,'-',1) as `lower`, substring_index(yourcolumn,'-',-1) as `upper`
from yourtable
This way you can return the records where a certain value falls between the range:
select * from yourtable
where 25 between substring_index(yourcolumn,'-',1) + 0 and substring_index(yourcolumn,'-',-1) + 0
The + 0 forces MySQL to convert the result of substring_index() to a numeric value before the comparison.
You can use the following solution using SUBSTRING_INDEX:
SELECT *
FROM table_name
WHERE 25 >= CONVERT(SUBSTRING_INDEX(SUBSTRING_INDEX(column_name, '-', 1), '-', -1), UNSIGNED INTEGER)
AND 25 <= CONVERT(SUBSTRING_INDEX(SUBSTRING_INDEX(column_name, '-', 2), '-', -1), UNSIGNED INTEGER)
-- or
SELECT *
FROM table_name
WHERE 25 BETWEEN CONVERT(SUBSTRING_INDEX(SUBSTRING_INDEX(column_name, '-', 1), '-', -1), UNSIGNED INTEGER)
AND CONVERT(SUBSTRING_INDEX(SUBSTRING_INDEX(column_name, '-', 2), '-', -1), UNSIGNED INTEGER)
demo: http://sqlfiddle.com/#!9/4ac7b3/3/0
I recommend you to change your table design. I would split the column using the VARCHAR datatype to two columns using the INTEGER datatype. You can add two new columns with the the following ALTER TABLE commands:
ALTER TABLE table_name ADD colNameA INT;
ALTER TABLE table_name ADD colNameB INT;
To split the values of you current column and update the values to the new columns you can use the following UPDATE command:
UPDATE table_name SET
colNameA = CONVERT(SUBSTRING_INDEX(SUBSTRING_INDEX(column_name, '-', 1), '-', -1), UNSIGNED INTEGER),
colNameB = CONVERT(SUBSTRING_INDEX(SUBSTRING_INDEX(column_name, '-', 2), '-', -1), UNSIGNED INTEGER)
At the end you can remove the VARCHAR column using this ALTER TABLE command:
ALTER TABLE table_name DROP COLUMN col_name
Now you can use the following (simple) query to get the expected results:
SELECT *
FROM table_name
WHERE 25 >= colNameA AND 25 <= colNameB
-- or
SELECT *
FROM table_name
WHERE 25 BETWEEN colNameA AND colNameB
If you want to get values beween 35 and 39, you can use below query,
SELECT
*
FROM
yourtable
WHERE
35 && 39
BETWEEN SUBSTRING_INDEX(tablecolumn, '-', 1) + 0 AND
SUBSTRING_INDEX(tablecolumn, '-', - 1) + 0
I don't know how it possible with MySQL.
But using php it possible to check with range.
For e.g.
// First of all get all record from database.
$search = 10; // Your searching value.
// Loop all rows.
while($rows = mysqli_fetch_array($r)){
$explode = explode("-",$rows['dash']); // For get from-to value.
$range = isset($explode[0])&&isset($explode[1])?range($explode[0],($explode[1]-1)):array(); // For get range.
if(in_array($search,$range)){ // For check searching value is exist or not !
echo "Yes ! I get into ".$rows['dash']; // Do stuff.
}
}
Note: If 10-15 then it will check with 10,11,12,13,14.
According to me if you dont want to change the table structure then,
Just fetch the records as per your other condition, Then from that data check your amount between that field using foreach loop and explode. like
If you have $data as all data
foreach($data as $value){
$new_val=explode(',',$value['new_field']);
if(25 >= $new_val[0] && 25 <= $new_val[1]){
// here create new array
}
}

How to replace a date with a standard date in SQL

Ok,
This is a little tricky, I am trying to replace the dates in a SQL Query results with a standard date, based on the month.
For example:
Any dates that are in July get 20140701
August gets 20140801
I could use a case statement:
Case
When Datepart(mm, TxnDate) = 1 and Datepart(yy, TxnDate) = 2014 then TxnDate = 20140101
etc...
but that could get very long as the database goes back 5 years and the result sets cover different periods then.
Any quick suggestions would be greatly appreciated.
Thanks,
declare #mydate datetime
select #mydate = GETDATE()
select cast(datepart(yy,#mydate) as varchar(4)) + RIGHT('0' + RTRIM(MONTH(#mydate)), 2) + '01'
select #mydate = GETDATE() - 10
select cast(datepart(yy,#mydate) as varchar(4)) + RIGHT('0' + RTRIM(MONTH(#mydate)), 2) + '01'
should print 20140701 and 20140601
If you simply want to set every TxnDate to the first of its month, you can do this:
TxnDate = DATEADD(month, DATEDIFF(month, 0, TxnDate), 0)
If your requirement is more complicated than that, you will need to explain.
Try Something like this
SELECT REPLACE (CONVERT(VARCHAR(8), TxnDate, 112),SUBSTRING(CONVERT(VARCHAR(8), TxnDate, 112),7,8),'01') AS [YYYYMMDD]
hope this is what you are looking for

Hundred year date conversion in ssis derived column

I'm trying to convert a Hundred Year Date (HYD) format to a regular date format through SSIS derived column transform. For example: Convert 41429
to 06/04/2013. I can do it with formatinng code within a script (and maybe I simply have to go this route) but feel there has to be a way to do so within a derived column that I'm just not getting. Any help is appreciated.
This is what I came up with. Are you sure your conversion is correct? My answer is 1 day. off.
DECLARE #t1 as date = '01/01/1900';
DECLARE #t2 as DATE = '12/31/1900';
DECLARE #hyd as INT;
-- This example shows that we need to add 1
SELECT #hyd = DATEDIFF (d, #t1, #t2) + 1 -- 364 + 1
SELECT #hyd
set #t2 = '06/04/2013';
SELECT #hyd = DATEDIFF (d,#t1, '06/04/2013') + 1-- 41427
SELECT #hyd
SELECT DATEADD (d, #hyd, '01-JAN-1900')
SELECT DATEADD (d, 41429, '01-JAN-1900')
A hundred year date is a calculation based on the number of days since 1899-12-31. It's an "Excel Thing". It also has a bug in it that you must account for.
The equivalent TSQL logic would be
DECLARE
#HYD int = 41429;
SELECT
#HYD =
CASE
WHEN #HYD > 60
THEN #HYD -1
ELSE
#HYD
END;
SELECT
DATEADD(d, #HYD, '1899-12-31') AS HYD;
Armed with that formula, I can write the following Expression in a Derived Column Transformation (assuming you have a column named HYD)
(HYD > 60) ? DATEADD("d",HYD - 1,(DT_DATE)"1899-12-31") : DATEADD("d",HYD,(DT_DATE)"1899-12-31")
And the results
--or inline SQL...using this
SELECT
case when ([HYD] > 60) then
DATEADD(day,[HYD] - 1,'1899-12-31')
else
DATEADD(day,[HYD],'1899-12-31')
end 'HYD_conv'
FROM
TableName
--and in the where clause if you like...
WHERE
(case when ([HYD] > 60) then DATEADD(day,[HYD] - 1,'1899-12-31') else DATEADD(day,[HYD],'1899-12-31') end) = '2016-01-14'

Convert seconds to human readable time duration

How would I best convert 90060 (seconds) to a string of "25h 1m"?
Currently I'm doing this in SQL:
SELECT
IF(
HOUR(
sec_to_time(
sum(time_to_sec(task_records.time_spent))
)
) > 0,
CONCAT(
HOUR(sec_to_time(sum(time_to_sec(task_records.time_spent)))),
'h ',
MINUTE(sec_to_time(sum(time_to_sec(task_records.time_spent)))),
'm'
),
CONCAT(
MINUTE(sec_to_time(sum(time_to_sec(task_records.time_spent)))),
'm'
)
) as time
FROM myTable;
But I'm not sure it's the most convenient method :-)
I'm open to suggestions on doing this both in SQL (differently than I'm already doing) or in PHP.
EDIT:
Examples of desired strings: "5m", "40m", "1h 35m", "45h" "46h 12m".
TIME_FORMAT(SEC_TO_TIME(task_records.time_spent),'%Hh %im')
Documentation is your friend:
http://dev.mysql.com/doc/refman/5.5/en/date-and-time-functions.html
According to comment:
DROP FUNCTION IF EXISTS GET_HOUR_MINUTES;
CREATE FUNCTION GET_HOUR_MINUTES(seconds INT)
RETURNS VARCHAR(16)
BEGIN
DECLARE result VARCHAR(16);
IF seconds >= 3600 THEN SET result = TIME_FORMAT(SEC_TO_TIME(seconds),'%kh %lm');
ELSE SET result = TIME_FORMAT(SEC_TO_TIME(seconds),'%lm');
RETURN result;
END
DELIMETER ;
Usage:
SELECT GET_HOUR_MINUTES(task_records.time_spent) FROM table
you can use predefined function sec_to_time()
sec_to_time(number_of_seconds)
http://dev.mysql.com/doc/refman/5.5/en/date-and-time-functions.html#function_sec-to-time
Try out this trick (on MySQL)
select date_format("1970-01-01 00:00:00" + interval <value_here> second, "%H:%i:%S")
SEC_TO_TIME sometime produces wrong values
try this: (input your pure seconds time)
var hours = Math.floor(input/3600);
var minutes = Math.floor((input-hours*3600)/60);
var seconds = input-(minutes*60)-(hours*3600);
function convertTime(){
return hours":"minutes":"seconds;
}

How to format int to price format in SQL?

I select the price 1000000 and I need to format it to $1,000,000. How can I do that in SQL?
To format with commas, you can use CONVERT with a style of 1:
declare #money money = 1000000
select '$' + convert(varchar, #money, 1)
will produce $1,000,000.00
If you want to remove the last 3 characters:
select '$' + left(convert(varchar, #money, 1), charindex('.', convert(varchar, #money, 1)) - 1)
and if you want to round rather than truncate:
select '$' + left(convert(varchar, #money + $0.50, 1), charindex('.', convert(varchar, #money, 1)) - 1)
Creating Function:
CREATE FUNCTION [dbo].[f_FormatMoneyValue]
(
#MoneyValue money
)
RETURNS VARCHAR(50)
AS
BEGIN
RETURN cast(#MoneyValue as numeric(36,2))
END
Using in Select Query:
Select dbo.f_FormatMoneyValue(isnull(SalesPrice,0))SalesPrice from SalesOrder
Output:
100.00
Formatting Money Value with '$' sign:
CREATE FUNCTION [dbo].[f_FormatMoneyWithDollar]
(
#MoneyValue money
)
RETURNS VARCHAR(50)
AS
BEGIN
RETURN '$' + convert(varchar, #MoneyValue, 1)
END
Output:
$100.00
Note: The above sample is for the money field. You can modify this function according to your needs
Hope this helps you..! :D
SELECT FORMAT(price, 'C2', 'en-us')
The SQL Server money datatype is just decimal(10, 4). To my knowledge there is no datatype that will present the way you want.
Adding the dollar sign and commas is something that should belong in the application logic, but if you really must do it through a database object consider adding the dollar sign, and commas every three characters (after the decimal point). In other words, you'll have to convert the int to varchar and do string manipulation.
It depends, however, there's no simple way to do it in standard SQL specs(SQL-92, SQL-2003, etc.).
For PostgreSQL PL/pgSQL and Oracle PL/SQL, you can use to_char to format numbers:
select to_char(1234567.123, 'FM$999,999,999.99')
Which gives output:
$1,234,567.12
See: http://www.postgresql.org/docs/7/static/functions2976.htm