I have a table date, subs, unsubs, where date is PK, and subs and unsubs are the number of subscribes and unsubscribes on that date.
I want to add a column to calculate the number of total subscribers on each day, that is, if sort date, subsall = subsall of last row + subs - unsubs.
For example: the table is
----------------
2012-01-01, 5, 2
2012-01-02, 6, 1
2012-01-03, 4, 3
----------------
I want to add a new column and the table becomes
-------------------
2012-01-01, 5, 2, 3
2012-01-02, 6, 1, 8
2012-01-03, 4, 3, 9
-------------------
What is the MySQL command to do that?
You are looking for a cumulative sum. Try this:
SELECT date, subs, unsubs, (#csum := #csum + subs - unsubs) cumulative
FROM table, (SELECT #csum :=0) c
ORDER BY date
You should create trigger on INSERT action, please see the manual: http://dev.mysql.com/doc/refman/5.0/en/create-trigger.html
You can just select the difference of the two columns and give it an alias of 'subsall' like this:
SELECT date, subs, unsubs, (subs - unsubs) as 'subsall' FROM table
SELECT (COUNT(subs) - COUNT(unsubs)) as total subscribers from table
GROUP BY DATE_FORMAT(date, '%d-%m-%Y')
Hope this works for you.
Related
I am trying to fetch the count of records entered in each month of the financial year
For example, I have declared a column called issue in varchar because the data what I am taking is issues of the particular machine. And for example, let's say one issue is raised in July month I enter the data as 'Jul 19-1' and the again issue is raised in the month of September again I go back to the issue happened in July and enter the data as 'sep19-2'.
So in the backend, it takes as jul19-1 sep19-2
What can be the query that I can write for counting the number of issues raised in each month
I tried the below query but
SELECT COUNT(month_nc)
FROM `ncr`
WHERE month_nc='Jul18-1'
In some months there will be only one issue so I can the count of the month given in the above query
What will be the query if I want to fetch the count of each month
id issue issue_month
1 bearing jul18-1
sep18-2
2 motor jul18-2
3 battery apr18-3
ps: issue_month is declared in varchar(10)
Here are two methods. One using strings:
select left(issue_month, 5), count(*)
from t
group by left(issue_month, 5), count(*)
This will not order the values correctly.
You can convert to a date to order properly:
order by str_to_date(concat('01', left(issue_month, 5)), '%d%b%y')
Or, represent the dates correctly:
select str_to_date(concat('01', left(issue_month, 5)), '%d%b%y') as yyyymm, count(*)
from t
group by yyyymm
order by yyyymm;
Here is what you can do to split your issue_month into "month_year" and "issue_count"
yourTable
select id,
issue,
issue_month,
REGEXP_SUBSTR(issue_month, '[^-]+', 1) as month_year,
REGEXP_SUBSTR(issue_month, '[^-]+', 1,2 ) as issue_count
from yourTable;
Now you can aggregate the issue_count across issues or year_months or any other field in your table.
For example, to get the sum of all the issues for any given month_year
select
month_year,
sum(issue_count) issue_count
from
(select
id, issue, issue_month,
REGEXP_SUBSTR(issue_month, '[^-]+', 1) as month_year,
REGEXP_SUBSTR(issue_month, '[^-]+', 1,2 ) as issue_count
from yourTable) foo
group by month_year;
I'm trying to allocate workers for a job for a specific date range and wanted to find out the minimum number of workers allocated for the given date range.
For example, my table contains
startDate endDate No.of.Workers
--------- --------- ---------------
1-1-2019 10-1-2019 1
11-1-2019 20-1-2019 1
now, i wanted to find out the minimum no of workers working in date range 1-1-2019 to 20-1-2019.
The output should be 1.
Suppose my table looks like,
startDate endDate No.of.Workers
--------- --------- ---------------
1-1-2019 10-1-2019 1
11-1-2019 20-1-2019 1
11-1-2019 15-1-2019 1
The output should be 2.
Is there any query for this in sql or i need to write an algorithm?
I am using mysql database.
You can get the number of workers needed by splitting the data, aggregating and using cumulative sums:
with dtes as (
select startDate as dte, numworks
from t
union all
select endDate as dte, - numworks
from t
)
select dte, sum(numworks),
sum(sum(numworks)) over (order by dte) as needed
from dtes
group by dte
order by dte;
To get the maximum, you can do something like this:
select dte, sum(numworks),
sum(sum(numworks)) over (order by dte) as needed
from dtes
group by dte
order by needed desc
fetch first 1 row only;
You don't specify the database, fetch first is ISO/ANSI standard SQL.
Also, it is not clear if the end date counts as one of the days. This can affect the results. If it is included, then you need to add one day to the "endDate" part of the logic. How you do that depends on your database.
I have a table - accounts, and it has 3 columns: id , timestamp and value.
I need to create a query that returns a table which in each row it will be the month (2, 3, etc.) and the sum of the values from this month.
for example: if my table will have 3 rows from January and one row from February, the query will return a two-rows table within the first row it'll be 1 and the sum of January's values, and the second will be 2 and the sum of February's values.
I have no idea how to begin. can anyone help me?
all you need to do is sum the value and also use mysql's MONTH() function to pull out the month from your timestamp
SELECT SUM(value) as total_amount, MONTH(timestamp) as month_num
FROM table
GROUP BY month_num
GROUP BY is used when you have an aggregate function (in our case its SUM) to know how to group your common fields. without a group by you will have all rows summed together
SQL Server can use the DATEPART function.
SELECT DATEPART(MONTH, your_date_column) AS month_
, SUM(your_value) AS value_
FROM your_table
GROUP BY DATEPART(MONTH, your_date_column)
I am trying to nest a few queries but so far am getting back error 1242: Subquery returns more than 1 row. I want more than one row, as I am working on a number of records.
I have 2 tables. One has a commencement date stored in 3 columns; yr_comm, mth_comm, day_comm. The 2nd table has a period of service (in years) for a number of users which is expressed as an integer (2.71, 3.45, etc).
I need to take this start date (from table 1), and add on the period of service (from table 2) to obtain an end date, but I only need to display the year.
I have 2 queries which work just fine when seperate, they result in the required values, however I am having trouble combining the queries to get the desired end result.
Query 1: Concatenate the 3 commencement values into date format
SELECT concat_ws('-', yr_comm, mth_comm, day_comm) AS date_comm
FROM table 1
Query 2: Convert the integer yrs_service into days
SELECT format(yrs_served * 365, 0) AS days_served
FROM table 2
Query 3: Use date_add function to add the days service to the commencement date
SELECT date_add(date_comm, INTERVAL days_served DAY) AS date_left
Can anyone suggest how I can achieve the above? Many thanks in advance.
EDIT - Here is the full query I am working on:
SELECT prime_minister.pm_name, yr_comm, party, ADDDATE(
(SELECT CONCAT_WS('-', yr_comm, mth_comm, day_comm) FROM ministry), INTERVAL
(SELECT FORMAT(yrs_served * 365, 0) FROM prime_minister) YEAR) AS date_left
FROM ministry JOIN prime_minister USING (pm_name)
WHERE party NOT LIKE '%labor%'
AND prime_minister.pm_name = ministry.pm_name
ORDER BY pm_name;
you can use user variables
SET #date = CONCAT_WS('-', 2012,1,1); -- paste your query here
SET #toAdd = (SELECT MONTH(CURDATE())); -- paste your query here
SELECT DATE_ADD(#date, INTERVAL #toAdd DAY) AS date_left
SQLFiddle Demo
which is the same as
SET #date = CONCAT_WS('-', 2012,1,1); -- paste your query here
SET #toAdd = (SELECT MONTH(CURDATE())); -- paste your query here
SELECT #date + INTERVAL #toAdd DAY AS date_left
SQLFiddle Demo
or without using variable, which is more longer,
SELECT (CONCAT_WS('-', 2012,1,1)) + INTERVAL (SELECT MONTH(CURDATE())) DAY AS date_left
SQLFiddle Demo
I'm looking to create the following chart from a MySQL database. I know how to actually create the chart (using excel or similar program), my problem is how to get the data needed to create the chart. In this example, I can see that on January 1, 60 tickets were in the state illustrated by the green line.
I need to track the historical state of tickets of a project through a date range. The date range is determined by a project manager (in this case it's January 1st through January 9th).
For each ticket, I have the following set of historical data. Each time something changes in the ticket (state, description, assignee, customer update, and other attributes not shown in this problem), a "timestamp" entry is made in the database.
ticket_num status_changed_date from_state to_state
123456 2011-01-01 18:03:44 -- 1
123456 2011-01-01 18:10:26 1 2
123456 2011-01-01 14:37:10 2 2
123456 2011-01-02 07:55:44 2 3
123456 2011-01-03 06:12:18 3 2
123456 2011-01-04 19:03:43 3 3
123456 2011-01-05 02:05:24 3 4
123456 2011-01-06 18:13:28 4 4
123456 2011-01-07 13:14:48 4 5
123456 2011-01-09 01:35:39 5 5
How can I query the database for a given time (determined by my script) and find out what state each of the tickets are in?
For example: To produce the chart shown above, given the date 2011-01-02 12:00:00, how many tickets were in the state "2"?
I've tried querying the database with specific dates and ranges, but can't figure out the proper way to get the data to create the chart. Thanks in advance for any help.
I'm not exactly sure I know what you want. But . . .
Assuming a table definition like:
create table ticket_data (ticket_num int,
status_changed_date datetime,
from_state int,
to_state int);
The following, for example would give you the number of values per day:
select date(status_changed_date) as status_date, count(*)
from ticket_data
group by status_date;
Now, if you want just from_state = 2, just add a where clause in to that effect. If you want just the ones on Jan 2, then add in where date(status_changed_date) = '2011-01-02'
Or, if you you're looking for the distinct number of tickets per day then, change count(*) to count(distinct ticket_num)
Is this what you're asking? SQL Fiddle here
Ok so if you are trying to get a count of records in a certain state at a certain time, I think a stored proc might be necessary.
CREATE PROCEDURE spStatesAtDate
#Date datetime,
#StateId int
AS
BEGIN
SET NOCOUNT ON;
SELECT COUNT(*) as Count
FROM ticket_table t1
WHERE to_state = #StateId AND status_changed_date < #Date
AND status_changed_date = (SELECT MAX(status_changed_date) FROM ticket_table t2 where t2.ticket_num=t1.ticket_num AND status_changed_date < #Date)
END
then to call this for the above example, you're query would look like
EXEC spStatesAtDate #Date='2011-01-02 12:00:00', #StateId=2
You can use a subquery to select the last modification date before a given point grouped by ticket_num and then select the states at this time.
SELECT
ticket_num,
to_state,
status_changed_date
FROM
tickets
WHERE
status_changed_date IN (
SELECT MAX(status_changed_date)
FROM tickets
WHERE status_changed_date < '2012-02-01 01:00:00'
GROUP BY ticket_num
)
It all boils down to common question: how to get list of items and their most recent statuses. So. Given one issue, we can get its most recent status with query:
select to_state
from ticket_states
where ticket_num = t.ticket_num
order by status_changed_date desc
limit 1
Next, we need to get all applicable distinct issue ids, which is a simple distinct select:
select distinct ticket_num from ticket_states
With these two subqueries we can already start building. For example, current list of issues and their latest statuses before specified date would be:
select t.ticket_num
, (select to_state
from ticket_states
where ticket_num = t.ticket_num
and status_changed_date <= '2012-01-01'
order by status_changed_date desc
limit 1) as last_state
from (select distinct ticket_num
from ticket_states) t;
All issues, which were non-existant at at the specified time will have last_state set to null.
This probably isn't the best way of doing this, but it is first which came to mind. I'll leave other stuff to you. Also I should mention that this is not a very efficient solution also.