Develop Reference

Determining Date Overlapping in a list of dates

I have a list of items that each have a start and end date time component.
var myDates= new List<Tuple<DateTime, DateTime>>();
Which I fill it out with some date.
Now I wanted to loop through them and see if any two of those have any overlapping date rang. So I did this:
var myOverlapList = (from start in myDates
from endDate in myDates
where !Equals(start, end)
where start.Item1 <= end.Item2 && start.Item2 >= end.Item1
select end);
It works when dates have overlap for example one day back and forth between two dates BUT it does NOT work when two date entries have the EXACT SAME values.
So how I can fix my code or just something else to achieve that.

The
where !Equals(startDate, endDate)
line, which is supposed to filter out the same date tuple actually is filtering out any duplicate, so any matching timespan falls out of the selection. So your query will return all DateTime tuples, which overlap with some other tuple in the collection, but only unique. And you want also to return tuples if they encounter in your collection more then once.Your problem, actually is that you can not differentiate between two different items with the same value. So you need a discriminator for them and because you use a list, the index of an item fits well. You can cast your Tuple<DateTime, DateTime> collection into, e. g. {int id, Tuple<DateTime, DateTime> range} object by
var datesWithId = dates.Select((d, i) => new {id = i, range = d});
and then modify your query like this:
var anyOverlap = (from startDate in datesWithId
from endDate in datesWithId
where startDate.id!=endDate.id
&& startDate.range.Item1 <= endDate.range.Item2
&& startDate.range.Item2 >= endDate.range.Item1
select endDate.range).Distinct();

How to filter list that has date and retrieve only data within 7 days

So I have this model:
Student Model
public int StudentId {get; set;}
public string StudentName {get; set;}
public DateTime EnrollDate {get; set;}
I also have a list of student Model which is something like
List<Student> listOfStudents = new List<Student>();
and inside that list there are 100 students detail and the enroll date.
What I do next is to sort the list into showing from the latest one to the oldest one.
listOfStudents.Sort((x, y) => DateTime.Compare(y.EnrollDate, x.EnrollDate));
and it's working. However, I am currently struggling in showing only the EnrollDate within 7 days from Today.

Conceptually, I think of LINQ a lot like SQL. You have the SELECT portion, which is your projection (i.e. what am I pulling out of this set of data?). If you omit the Select() clause from LINQ, you'll get the whole record vs. only a portion if you wanted to pluck out only pieces of it. You have your WHERE portion which is a limiter, or filter condition that when applied to the set pulls back only the records that satisfy said condition. And lastly, there are operations you can apply that affect the order of the returned set. That's where the OrderBy() and OrderByDescending() come into play. So lets map those concepts to the examples below
No Select(), but we do have a Where() and an OrderBy()
var then = DateTime.Now.AddDays(-7); //One portion of our Where. More below
var sortedStudents = listOfStudents
//Our predicate. 's' = the Student passed to the function. Give me only the students
//where s.EnrollDate is greater or equal to the variable 'then' (defined above)
.Where(s => s.EnrollDate >= then)
//We have no Select statement, so return whole students
//And order them by their enrollment date in ascending order
.OrderBy(s => s.EnrollDate);
When run, sortedStudents will be loaded up only with students (entire Student objects, not a projection) that meet our Where() criteria. The Where() function takes predicate that specifies our criteria. A predicate is simply a function that accepts a record from the set that we're filtering, and returns a bool indicating whether or not it should be included.
Let's change the filter by adjusting the Where()
//Notice we've changed 'then' from 7 days ago to a fixed point in time: 26 June 2018
var then = new DateTime.Parse("26 June 2018");
var sortedStudents = listOfStudents
.Where(s => s.EnrollDate >= then)
//Still no Select(). We'll do that next
.OrderBy(s => s.EnrollDate);
Just like before sortedStudents will have whole Student records, but this time it will only contain those enrolled after or on 26 June 2018, as specified by our predicate.
Let's add a Select()
var then = new DateTime.Parse("26 June 2018");
var dates = listOfStudents
.Where(s => s.EnrollDate >= then)
.Select(s => s.EnrollDate);
Now we've changed it so that instead of pulling back a whole Student we're only plucking out the EnrollDate. Notice I've changed the name of the receiving variable from sortedStudents to dates reflecting the fact that it now only contains a list of DateTime objects.
You could still replace .OrderBy() with .OrderByDescending() to change the order.

How about breaking down the problem into 2 sub-problems?
Sub-problem #1
showing only the EnrollDate within 7 days from Today
We only need Students whose EnrollDate property is within 7 days from today:
var today = DateTime.UtcNow;
sevenDaysOldList = listOfStudents.Where(x => (today - x.EnrollDate).TotalDays < 7);
The subtraction of the two dates results in a TimeSpan with a TotalDays property, which we can use to determine the number of days elapsed between the two dates.
Sub-problem #2
sort the list into showing from the latest one to the oldest one.
We need to sort sevenDaysOldList by EnrollDate in descending order:
sevenDaysOldList.Sort((x, y) => y.EnrollDate.CompareTo(x.EnrollDate));
..which will sort the list in place. OrderByDescending is a good candidate for this (it returns a new ordered list implementing IOrderedEnumerable<T>):
sevenDaysOldList.OrderByDescending(x => x.EnrollDate);
// and of course .OrderBy(x => x.EnrollDate) for ascending order
Combine #1 & #2
You can now combine the solutions of the two sub-problems into one. How you do it is at your own discretion. This is how I would do it:
var sevenDaysOldList = listOfStudents.Where(x => (today - x.EnrollDate).TotalDays < 7)
.OrderByDescending(x => x.EnrollDate);
Update: question in comment
How do I modify/sort the list that remove all the list less than "26 June 2018" ? So the list will only have data date greater than 26 June 2018. Any data with date before 26 June will be removed
You can initialize that date in a DateTime variable, and use it with List<T>.RemoveAll(Predicate<T>), to remove items in sevenDaysOldList which are smaller than that date:
var filterDate = new DateTime(2018, 06, 26);
sevenDaysOldList.RemoveAll(x => x.EnrollDate < filterDate);

LINQ: Group by date including the year

I have dataset looks like this:
FileName Date
ABC - 01/10/16
DBC - 01/11/16
ZYX - 03/10/16
ABX2 - 01/10/17
IOS - 01/09/17
How can I group them into a list of groups of months while ensuring that the year is taken into account in the clause?
I'm currently using a LINQ Query is creating groups by month but not including the year, so I have a group of ABC, ZYX and ABX2. even though ABX2 was a 2017 report but the same month so should be in a different group.
I've been trying different ways of doing this but none of have been successful as of yet.
var newList = from x in list
group x
by x.Properties.LastModified.Value.Month into lastMod
where lastMod.Count() > 1
select lastMod;
Once I have them in separate groups, I will find out which one was written last and save that and remove the rest. I'm quite stuck and been on this issue for half day. would appreciate fresh eyes on it.

You can group by a composite year-month key, like this:
var newList = list
.Where(x => x.Properties.LastModified.HasValue)
.GroupBy(x => new {
x.Properties.LastModified.Value.Month
, x.Properties.LastModified.Value.Year
})
.Where(g => g.Count() > 1);
You need to ensure that LastModified has non-null value before accessing its Value property.

I can't test this at the moment but I think grouping by an anonymous type that contains both the month and year should do it.
var newList = from x in list
group x
by new {x.Properties.LastModified.Value.Month, x.Properties.LastModified.Value.Year} into lastMod
where lastMod.Count() > 1
select lastMod;

Select if criteria is met

Lets say I have a datatable with three columns: timestamp, curveID and price. I would like to give a time and then select for each day the timestamp, curveID and price but only if all curveIDs are present.
The problem is, not for every time all the data is present, so at 10:00:00 there might be only data for curveID 1 but nothing for ID =2, and so forth.
I thought I could do the following to select the first dataset where all curveIDs are there and time is greater or equal to my criteria:
dataSet.ReadXml(#"C:\temp\Prices.xml", XmlReadMode.InferTypedSchema);
ds = dataSet.Tables[0];
var dt = ds.Clone();
int criteria = 10;
var list = ds.AsEnumerable().Where(x => x.Field<DateTime>("Timestamp").Hour >= criteria)
.GroupBy(x => new{Date = x.Field<DateTime>("Timestamp").Date, Curve = x.Field<object>("CurveID")})
.First().ToList();
However, this returns multiple records on the same day (at different times) for the same curve ID.
I would like to return only a single record for each curveID on each day at a time close to the criteria time where all curveIDs are present.
For clarity, lets say I m looking for curveID 1 & 2, if at 10:00:00 on day 1 only curveID 1 is present but curveID 2 is missing I would need to check whether at 10:01:00 both are there, if yes I take for that day the two record sets from that time. This I would have to check for every day in the database

// criteria is your integer Hour representation
var criteria = 10;
// array of curveIds to look for
var curveIds = new int[] {1, 2};
var result =
// grouping by date first
ds.GroupBy(x => x.Field<DateTime>("Timestamp").Date,
(date, items) => new { date, items = items
// items with the same timestamp go to one group
.GroupBy(i => i.Field<DateTime>("Timestamp"), (datetime, timestampItems) => new { datetime, timestampItems })
// filter by criteria
.Where(dti => dti.datetime.Hour >= criteria)
// filter by curveIds
.Where(dti => curveIds.All(cid => dti.timestampItems.Any(tsi => tsi.Field<int>("curveID") == cid)))
.OrderBy(dti => dti.datetime)
.FirstOrDefault() });
In the end you will receive a "per day" result fitting all your mentioned requirements: occurs after some criteria, have all curveIds, be earliest one.

You may want to group by Date first and then by hour using something like
group thing by new {
firstThing = x.Field<DateTime>("TimeStamp").Date,
secondThing = x.Field<DateTime>("TimeStamp").Date.Hour,
}
My syntax is probably off by a little, but that should get you moving in the right direction

How can I group a date column to a less precise format while selecting in Entity Framework or LINQ?

I am collecting data within every ten seconds (six records in a minute).
Using Entity Framework or LINQ, I want to obtain the average of the records within a every minute.
Simply, Date column is in (%Y.%m.%d %H:%i:%s) format, and i want to group by (%Y.%m.%d %H:%i) format in mysql using Entity Framework or LINQ

Assuming that you mean you have a date time column you can group by the DateTime properties Year, Month, Day, Hour, and Minute.
var results = from row in db.SomeTable
group by new
{
row.Date.Year,
row.Date.Month,
row.Date.Day,
row.Date.Hour,
row.Date.Minute
} into grp
select new
{
YMDHM = grp.Key
SomeAverage = grp.Average(x => x.SomeValueToAverage)
};
Then when you iterate the results you can turn the values in YMDHM back into a DateTime. Here's an example where you could turn the results into a Dictionary.
var dictionaryOfAverages = results.ToDictionary(
x => new DateTime(x.YMDHM.Year,
x.YMDHM.Month,
x.YMDHM.Day,
x.YMDHM.Hour,
x.YMDHM.Minute,
0),
x => x.SomeAverage);
On the other hand if you actually mean you are storing the date time in the DB as a formatted string then the following would be what you want
var results = from row in db.SomeTable
group by new row.Date.SubString(0, 16) into grp
select new
{
YMDHM = grp.Key
SomeAverage = grp.Average(x => x.SomeValueToAverage)
};
This is assuming that your string formatted dates look like "2013.05.08 07:25:33" If the format isn't fixed width with leading zeros for month, day, and/or hour you'd have to do something like row.Date.SubString(0, row.Date.Length - 3) instead