Develop Reference

how to check if dates exist in a list of all the weeks of a year

I've got an ASP.net C# application which creates a list of all the weeks in a given year. e.g. selected year 2019, and will produce 31/12/2019 to 06/01/2019 and so on. see attached image.
To produce this I am borrowing some code from an example I found on Stack Overflow here
Now I also have another list containing dates in the format dd/MM/yyyy, this is generated from an XML file, so I wanted to only show the weeks that match dates in the weeks of the year list and populate the drop down list when a date in my XML generated list is contained within it.
For example if I had a full week or even a day in my XML generated list which fell between the 31/12/2018 to 06/01/2019 I want to show it in the drop down list.
Similarly if the XML generated list doesn't contain at least a day from that week then don't show it.
I've pasted the code I used to get the weeks of a given year below.
I'm not sure of any easy way to compare both lists. Any help would be greatly appreciated.
public List<string> FetchWeeks(int year)
{
List<string> weeks = new List<string>();
DateTime startDate = new DateTime(year, 1, 1);
startDate = startDate.AddDays(1 - (int)startDate.DayOfWeek);
DateTime endDate = startDate.AddDays(6);
while (startDate.Year < 1 + year)
{
weeks.Add(string.Format("{0:dd/MM/yyyy} to {1:dd/MM/yyyy}", startDate, endDate));
startDate = startDate.AddDays(7);
endDate = endDate.AddDays(7);
}
//DropDownList1.Items.Add(weeks);
return weeks;
}

If were trying to compare lists to determine a set of valid weeks, I would try to determine an absolute week index and use that in my work. Since weeks are not impacted by things like leap years or other date oddities, we can just count in 7-day intervals from the beginning of a known of date range. Forgive me if my C# is rusty, but something to the effect of:
public int ToWeekIndex(DateTime date)
{
// Takes any date and maps it to a value that represents the week it resides in.
Timespan ts = date - DateTime.MinValue // Monday, January 1, 0001;
return ts.Days / 7; // Integer divide, drops the remainder.
}
public DateTime FromWeekIndex(int weekIndex)
{
// Takes a week index and returns the Monday from it.
Timespan ts = new Timespan(weekIndex * 7, 0, 0, 0); // Days, hours, minutes, seconds
return DateTime.MinValue + ts;
}
Then to build out your weeks, you could do something to the effect of pseudocode:
all_weeks = []
for date in January 1, 2019 to December 31, 2019 step 7 days:
week_index = ToWeekIndex(date)
week_start = FromWeekIndex(week_index)
week_end = week_start + 7 days - 1 second
all_weeks += [week_start, week_end]

instead of a list of string for dates, use a list of object that contains the week dates and a boolean defaulted to false.
public class WeekObj
{
public string Week { get; set; }
public DateTime StartDate { get; set; }
public DateTime EndDate { get; set; }
public bool IsValid { get; set; }
};
List<WeekObj> weeks= new WeekObj();
weeks.add(new WeekObj { "week string", startDate, endDate, false });
Loop through your list of days, and for each day go through the list of weeks and set it to true if the day is between startDate and endDate (which is now in DateTime obj), do that for false dates, no need to recompare true dates.
public static bool Between(DateTime input, DateTime date1, DateTime date2)
{
return (input >= date1 && input <= date2);
}

This is the weeks in the Year data you already have -
class WeekData
{
public DateTime WeekStartDate { get; set; }
public DateTime WeekEndDate { get; set; }
public int WeekStartDay //Gets Day in the year for the Week Start Date
{
get { return WeekStartDate.DayOfYear; }
}
public int WeekEndDay //Gets Day in the year for the Week End Date
{
get { return WeekEndDate.DayOfYear; }
}
}
Dummy WeeksInTheYear data
List<WeekData> weeks = new List<WeekData>
{
new WeekData{WeekStartDate = new DateTime(2019,10,6), WeekEndDate = new DateTime(2019,10,12)},
new WeekData{WeekStartDate = new DateTime(2019,10,13), WeekEndDate = new DateTime(2019,10,19)},
new WeekData{WeekStartDate = new DateTime(2019,10,20), WeekEndDate = new DateTime(2019,10,26)},
new WeekData{WeekStartDate = new DateTime(2019,10,27), WeekEndDate = new DateTime(2019,11,2)}
};
Dummy Dates from the XML feed
List<DateTime> xmlDates = new List<DateTime> { new DateTime(2019, 11, 1), new DateTime(2019, 10, 12), new DateTime(2019, 10, 31) };
Filtering
var weeksINeed = new List<WeekData>();
foreach (var date in xmlDates)
{
var weekINeed = weeks.Where(x => x.WeekStartDay <= date.DayOfYear && x.WeekEndDay >= date.DayOfYear)
.FirstOrDefault();
if (!weeksINeed.Any(x => x.WeekStartDay == weekINeed.WeekStartDay))
{
weeksINeed.Add(weekINeed);
}
}
Output -
foreach (var weekdata in weeksINeed.OrderBy(x=>x.WeekStartDay))
{
Console.WriteLine($"WeekStartDate - {weekdata.WeekStartDate} WeekEndDate - {weekdata.WeekEndDate}");
}

Using some extension functions and LINQ, you can just generate the list directly from the XML Date List<string>.
First, an IEnumerable<> extension to select distinct by a lambda function:
public static class IEnumerableExt {
public static IEnumerable<T> DistinctBy<T, TKey>(this IEnumerable<T> src, Func<T, TKey> keySelector, IEqualityComparer<TKey> comparer = null) {
var seenKeys = new HashSet<TKey>(comparer);
foreach (var e in src)
if (seenKeys.Add(keySelector(e)))
yield return e;
}
}
Then some calendar extensions using the built-in ISOWeek methods to get the week of year (Based on your week date ranges, I assume you are using ISO 8601 Weeks):
public static class CalendarExt {
public static int GetISO8601WeekOfYear(this DateTime aDate) => ISOWeek.GetWeekOfYear(aDate);
public static DateTime FirstDateOfYear(this DateTime d) => new DateTime(d.Year, 1, 1);
public static DateTime FirstDateOfISO8601Week(this DateTime aDate) => aDate.AddDays(-(((int)aDate.DayOfWeek + 6) % 7));
public static DateTime LastDateofISO8601Week(this DateTime aDate) => aDate.FirstDateOfISO8601Week().AddDays(6);
public static DateTime FirstDateOfISO8601Week(int year, int weekNum) => ISOWeek.ToDateTime(year, weekNum, DayOfWeek.Monday);
public static DateTime LastDateofISO8601Week(int year, int weekNum) => FirstDateOfISO8601Week(year, weekNum).AddDays(6);
// for .Net without ISOWeek
//public static DateTime FirstDateOfISO8601Week(this DateTime aDate) => aDate.AddDays(-(((int)aDate.DayOfWeek + 6) % 7));
//public static int GetISO8601WeekOfYear(this DateTime aDate) =>
// CultureInfo.CurrentCulture.Calendar.GetWeekOfYear(aDate.AddDays(DayOfWeek.Monday <= aDate.DayOfWeek && aDate.DayOfWeek <= DayOfWeek.Wednesday ? 3 : 0), CalendarWeekRule.FirstFourDayWeek, DayOfWeek.Monday);
}
Finally, given your list of string dates from XML in xmlDateStrings, you can compute the week ranges list:
var currentCulture = CultureInfo.CurrentCulture;
var ans = xmlDateStrings.Select(ds => DateTime.ParseExact(ds, "dd/MM/yyyy", currentCulture))
.DistinctBy(d => d.GetISO8601WeekOfYear())
.OrderBy(d => d) // assume XML is unsorted
.Select(d => $"{d.FirstDateOfISO8601Week().ToString("dd/MM/yyyy")} to {d.LastDateofISO8601Week().ToString("dd/MM/yyyy")}")
.ToList();

Perform calculations on a set using only LINQ

I'd like to perform the following using only LINQ.
I have a list of time sheet entries with user's in and out times. The class looks like this:
public class TimeSheetLog
{
public Guid EmployeeId { get; set; }
public DateTime ClockInTimeStamp { get; set; }
public DateTime ClockOutTimeStamp { get; set; }
}
I'm passing a List<TimeSheetLog>() which contains all logs from the beginning of the year to date.
I'm trying to calculate the total work time -- regardless of employee -- for the month of January. Please also notice that I have a function named GetTimeDifferenceInMinutes() which calculates the number of minutes between two date/time values.
Here's what I currently have but I feel the whole thing can be done using LINQ only.
public static int GetTotalTimeWorked(List<TimeSheetLog> logs, DateTime startDate, DateTime endDate)
{
// I'm passing 1/1/2018 for startDate and 1/31/2018 for endDate to this function
var totalTimeWorkedInMinutes = 0;
var januaryLogs = logs.Where(x => x.ClockInTimeStamp >= startDate &&
x.ClockOutTimeStamp <= endDate);
foreach(var item in januaryLogs)
{
totalTimeWorkedInMinutes += GetTimeDifferenceInMinutes(item.ClockInTimeStamp, itemClockOutTimeStamp);
}
return totalTimeWorkedInMinutes;
}

var logsFilteredByDate = logs.Where(x => x.ClockInTimeStamp >= startDate &&
x.ClockOutTimeStamp <= endDate);
var totalTimeWorkedInMinutes = logsFilteredByDate.Sum(x =>
GetTimeDifferenceInMinutes(x.ClockInTimeStamp, x.ClockOutTimeStamp));
Or, to combine it all into one query, which is unnecessary and harder to read,
var totalTimeWorkedInMinutes = logs.Where(x => x.ClockInTimeStamp >= startDate &&
x.ClockOutTimeStamp <= endDate)
.Sum(x =>
GetTimeDifferenceInMinutes(x.ClockInTimeStamp, x.ClockOutTimeStamp));

you need sum
var tot = januaryLogs.Sum(item=>GetTimeDifferenceInMinutes(item.ClockInTimeStamp, itemClockOutTimeStamp));

Couldn't you do the Where with a Sum and do DateTime Subtract in the Sum, so
decimal total = logs.Where(x => x.ClockInTimeStamp >= startDate && x.ClockOutTimeStamp <= endDate).Sum(x.ClockOutTimeStamp.Subtract(x.ClockInTimeStamp).TotalMinutes);

The problem seems easy until you realize that a time sheet can span months. So if someone clocked in on January 31st and clocked out on February 1st, you have to count partial timesheets, to do it right.
Here is my solution:
public static class ExtensionMethods
{
static public double TotalMinutes(this IEnumerable<TimeSheetLog> input, DateTime startPeriod, DateTime endPeriod)
{
return TimeSpan.FromTicks
(
input
.Where( a=>
a.ClockOutTimeStamp >= startPeriod &&
a.ClockInTimeStamp <= endPeriod
)
.Select( a=>
Math.Min(a.ClockOutTimeStamp.Ticks, endPeriod.Ticks) -
Math.Max(a.ClockInTimeStamp.Ticks, startPeriod.Ticks)
)
.Sum()
)
.TotalMinutes;
}
}
Logic:
Find all timesheets that overlap at least partially with the period of interest.
Compute the start time as either the clock in time or the period start time, whichever is later.
Compute the end time as either the clock out time or the period end time, whichever is earlier.
Take the difference of the start and end time as ticks. Sum() these.
To do all this math, we convert all the timestamps to Ticks, since you can't take a Max() of two DateTimes. We can add ticks up just fine, then convert the total back into minutes before returning.
Test program (notice the third timesheet spans both January and February):
public class Program
{
static public List<TimeSheetLog> testData = new List<TimeSheetLog>
{
new TimeSheetLog
{
ClockInTimeStamp = DateTime.Parse("1/1/2018 9:00 am"),
ClockOutTimeStamp = DateTime.Parse("1/1/2018 5:00 pm")
},
new TimeSheetLog
{
ClockInTimeStamp = DateTime.Parse("1/2/2018 9:00 am"),
ClockOutTimeStamp = DateTime.Parse("1/2/2018 5:00 pm")
},
new TimeSheetLog
{
ClockInTimeStamp = DateTime.Parse("1/31/2018 6:00 pm"),
ClockOutTimeStamp = DateTime.Parse("2/1/2018 9:00 am")
},
new TimeSheetLog
{
ClockInTimeStamp = DateTime.Parse("2/3/2018 9:00 am"),
ClockOutTimeStamp = DateTime.Parse("2/3/2018 5:00 pm")
}
};
public static void Main()
{
var startPeriod = new DateTime(2018, 1, 1);
var endPeriod = new DateTime(2018, 1, 31, 23, 59, 59, 9999);
Console.WriteLine( testData.TotalMinutes(startPeriod, endPeriod).ToString("0.00") );
}
}
Output:
1320.00
...which is correct.
See my code on DotNetFiddle

Another option is to use .Aggregate function.
public static int GetTotalTimeWorked(List<TimeSheetLog> logs, DateTime startDate, DateTime endDate)
{
var totalTimeWorkedInMinutes = 0;
return logs.Where(x => x.ClockInTimeStamp >= startDate && x.ClockOutTimeStamp <= endDate)
.Aggregate(totalTimeWorkedInMinutes, (total, item) => total + GetTimeDifferenceInMinutes(item.ClockInTimeStamp, item.ClockOutTimeStamp));
}

ASP.NET MVC Filter datetime by weeks

I've got a Web API and a Get method, returning a query:
var query = from results in context.Table
where results.Date>= startDate && results.Date <= endDate
select new
{
Week = { this is where I need a method to group by weeks },
Average = results.Where(x => x.Number).Average()
}
return query.ToList();
I want to calculate the average for each 7 days (that being the first week).
Example:
Average 1 ... day 7 (Week 1)
Average 2 ... day 14 (Week 2)
How can I do that? Being given an interval of datetimes, to filter it by weeks (not week of year)

Try this (not tested with tables)
var avgResult = context.QuestionaireResults
.Where(r => (r.DepartureDate >= startDate && r.DepartureDate <= endDate)).ToList()
.GroupBy( g => (Decimal.Round(g.DepartureDate.Day / 7)+1))
.Select( g => new
{
Week = g.Key,
Avg = g.Average(n => n.Number)
});

You will need to group by the number of days, since a reference date, divided by 7, so
.GroupBy(x => Math.Floor(((x.DepartureDate - new DateTime(1980,1,1)).TotalDays + 2) / 7))
Subtracting "Jan 1, 1980" from your departure date, gives you a TimeSpan object with the difference between the two dates. The TotalDays property of that timespan gives you timespan in days. Adding 2 corrects for the fact that "Jan 1, 1980" was a Tuesday. Dividing by 7 gives you the number of weeks since then. Math.Floor rounds it down, so that you get a consistent integer for the week, given any day of the week or portion of days within the week.
You could simplify a little by picking a reference date that is a Sunday (assuming that is your "first day of the week"), so you dont have to add 2 to correct. Like so:
.GroupBy(x => Math.Floor(((x.DepartureDate - new DateTime(1979,12,30)).TotalDays) / 7))
If you are sure that your data all falls within a single calendar year, you could maybe use the Calendar.GetWeekOfYear method to figure out the week, but I am not sure it would be any simpler.

Why not write a stored procedure, I think there may be some limitations on your flexibility using Linq because of the idea that normally the GroupBy groups by value (the value of the referenced "thing") so you can group by State, or Age, but I guess you can Group week... (new thought)
Add a property called EndOfWeek and for example, the end of this week is (Sunday let's say) then EndOfWeek = 9.2.16 whereas last week was 8.28.16... etc. then you can easily group but you still have to arrange the data.
I know I didn't answer the question but I hope that I sparked some brain activity in an area that allows you to solve the problem.
--------- UPDATED ----------------
simple solution, loop through your records, foreach record determine the EndOfWeek for that record. After this you will now have a groupable value. Easily group by EndOfWeek. Simple!!!!!!!!!!!! Now, #MikeMcCaughan please tell me how this doesn't work? Is it illogical to extend an object? What are you talking about?
------------ HERE IS THE CODE ----------------
using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Threading.Tasks;
namespace SandboxConsole
{
class Program
{
static void Main(string[] args)
{
var t = new Transactions();
List<Transactions> transactions = t.GetTransactions();
// Now let's add a Weeks end date so we can determine the average per week
foreach(var transaction in transactions)
{
var transactionDayOfWeek = transaction.TransactionDate;
int daysUntilEndOfWeek_Sat = ((int)DayOfWeek.Saturday - (int)transactionDayOfWeek.DayOfWeek + 7) % 7;
transaction.Newly_Added_Property_To_Group_By_Week_To_Get_Averages = transactionDayOfWeek.AddDays(daysUntilEndOfWeek_Sat).ToShortDateString();
//Console.WriteLine("{0} {")
}
foreach(var weekEnd in transactions.GroupBy(tt => tt.Newly_Added_Property_To_Group_By_Week_To_Get_Averages))
{
decimal weekTotal = 0;
foreach(var trans in weekEnd)
{
weekTotal += trans.Amount;
}
var weekAverage = weekTotal / 7;
Console.WriteLine("Week End: {0} - Avg {1}", weekEnd.Key.ToString(), weekAverage.ToString("C"));
}
Console.ReadKey();
}
}
class Transactions
{
public int Id { get; set; }
public string SomeOtherProp { get; set; }
public DateTime TransactionDate { get; set; }
public decimal Amount { get; set; }
public string Newly_Added_Property_To_Group_By_Week_To_Get_Averages { get; set; }
public List<Transactions> GetTransactions()
{
var results = new List<Transactions>();
for(var i = 0; i<100; i++)
{
results.Add(new Transactions
{
Id = i,
SomeOtherProp = "Customer " + i.ToString(),
TransactionDate = GetRandomDate(i),
Amount = GetRandomAmount()
});
}
return results;
}
public DateTime GetRandomDate(int i)
{
Random gen = new Random();
DateTime startTime = new DateTime(2016, 1, 1);
int range = (DateTime.Today - startTime).Days + i;
return startTime.AddDays(gen.Next(range));
}
public int GetRandomAmount()
{
Random rnd = new Random();
int amount = rnd.Next(1000, 10000);
return amount;
}
}
}
------------ OUTPUT ---------------
Sample Output

How to get gap in date ranges from a period of time

I have an initial and a final date range = 1/1/2015 - 1/30/2015
I have these date ranges that represent dates of unavailability.
1/5/2015 - 1/10/2015
1/15/2015 - 1/20/2015
1/22/2015 - 1/28/2015
I want this output, mainly the dates of availability from the main range:
A: 1/1/2015 - 1/4/2015
B: 1/11/2015 - 1/14/2015
C: 1/21/2015 - 1/21/2015
D: 1/29/2015 - 1/30/2015
I tried to generate a sequential date range like this in order to get the exception dates with Except() but I think I'm complicating the thing.
//dtStartDate = 1/1/2015
//dtEndDate = 1/30/2015
var days = (int)(dtEndDate - dtStartDate).TotalDays + 1;
var completeSeq = Enumerable.Range(0, days).Select(x => dtStartDate.AddDays(x)).ToArray();
How can I get the gap of date ranges from period of time.
I other words how can I get the A, B, C and D from this picture
http://www.tiikoni.com/tis/view/?id=ebe851c
If these dates overlap, they must not be considered only where is a gap.
----------UPDATE-----------
I think if I do this:
var range = Enumerable.Range(0, (int)(1/10/2015 - 1/5/2015).TotalDays + 1).Select(i => 1/5/2015.AddDays(i));
var missing = completeSeq.Except(range).ToArray();
for each date range I will have the exclusion of each date range given but still cannot get the gap!

I saw your question in my morning today and really liked it, but was busy the whole day. So, got a chance to play with your question and believe me I enjoyed it. Here is my code:-
DateTime startDate = new DateTime(2015, 1, 1);
DateTime endDate = new DateTime(2015, 1, 30);
int totalDays = (int)(endDate - startDate).TotalDays + 1;
availability.Add(new Availability { StartDate = endDate, EndDate = endDate });
var result = from x in Enumerable.Range(0, totalDays)
let d = startDate.AddDays(x)
from a in availability.Select((v, i) => new { Value = v, Index = i })
where (a.Index == availability.Count - 1 ?
d <= a.Value.StartDate : d < a.Value.StartDate)
&& (a.Index != 0 ? d > availability[a.Index - 1].EndDate : true)
group new { d, a } by a.Value.StartDate into g
select new
{
AvailableDates = String.Format("{0} - {1}",g.Min(x => x.d),
g.Max(x => x.d))
};
This, definitely need explanation so here it is:-
Step 1: Create a range of dates from Jan 01 till Jan 30 using Enumerable.Range
Step 2: Since after the second unavailable date range, we need to limit the dates selected from last endate till current object startdate, I have calculated index so that we can get access to the last enddate.
Step 3: Once we get the index, all we need to do is filter the dates except for first date range since we didn't have last object in this case.
Step 4: For the last item since we don't have the max range I am adding the endDate to our unavailable list (hope this makes sense).
Here is the Working Fiddle, if you get confused just remove group by and other filters and debug and see the resulting output it will look fairly easy :)

using System;
using System.Collections.Generic;
using System.Linq;
public static class Program {
public static void Main() {
Tuple<DateTime,DateTime> range=Tuple.Create(new DateTime(2015,1,1),new DateTime(2015,1,30));
Tuple<DateTime,DateTime>[] exclude=new[] {
Tuple.Create(new DateTime(2015,1,5),new DateTime(2015,1,10)),
Tuple.Create(new DateTime(2015,1,15),new DateTime(2015,1,20)),
Tuple.Create(new DateTime(2015,1,22),new DateTime(2015,1,28))
};
foreach(Tuple<DateTime,DateTime> r in ExcludeIntervals(range,exclude)) {
Console.WriteLine("{0} - {1}",r.Item1,r.Item2);
}
}
public static IEnumerable<Tuple<DateTime,DateTime>> ExcludeIntervals(Tuple<DateTime,DateTime> range,IEnumerable<Tuple<DateTime,DateTime>> exclude) {
IEnumerable<Tuple<DateTime,bool>> dates=
new[] { Tuple.Create(range.Item1.AddDays(-1),true),Tuple.Create(range.Item2.AddDays(1),false) }.
Concat(exclude.SelectMany(r => new[] { Tuple.Create(r.Item1,false),Tuple.Create(r.Item2,true) })).
OrderBy(d => d.Item1).ThenBy(d => d.Item2); //Get ordered list of time points where availability can change.
DateTime firstFreeDate=default(DateTime);
int count=1; //Count of unavailability intervals what is currently active. Start from 1 to threat as unavailable before range starts.
foreach(Tuple<DateTime,bool> date in dates) {
if(date.Item2) { //false - start of unavailability interval. true - end of unavailability interval.
if(--count==0) { //Become available.
firstFreeDate=date.Item1.AddDays(1);
}
} else {
if(++count==1) { //Become unavailable.
DateTime lastFreeDate=date.Item1.AddDays(-1);
if(lastFreeDate>=firstFreeDate) { //If next unavailability starts right after previous ended, then no gap.
yield return Tuple.Create(firstFreeDate,lastFreeDate);
}
}
}
}
}
}
ideone.com

Got a little oopy...
public class DateRange
{
public DateTime Start { get; set; }
public DateTime End { get; set; }
public bool HasStart
{
get { return Start != DateTime.MinValue; }
}
public bool IsInRange(DateTime date)
{
return (date >= this.Start && date <= this.End);
}
public List<DateRange> GetAvailableDates(DateRange excludedRange)
{
return GetAvailableDates(new List<DateRange>(){excludedRange});
}
public List<DateRange> GetAvailableDates(List<DateRange> excludedRanges)
{
if (excludedRanges == null)
{
return new List<DateRange>() { this };
}
var list = new List<DateRange>();
var aRange = new DateRange();
var date = this.Start;
while (date <= this.End)
{
bool isInARange = excludedRanges.Any(er => er.HasStart && er.IsInRange(date));
if (!isInARange)
{
if (!aRange.HasStart)
{
aRange.Start = date;
}
aRange.End = date;
}
else
{
if (aRange.HasStart)
{
list.Add(aRange);
aRange = new DateRange();
}
}
date = date.AddDays(1);
}
if (aRange.HasStart)
{
list.Add(aRange);
}
return list;
}
}

Parallel.For with date time

OK this code is a bit meta but it roughly explains how i have it now and what i want to achieve.
specialObject{
DateTime date;
int number;
}
var startDate = Lowest date in the list;
var endDate = Hightest date int the list;
List<SpecialObject> objs = (list from database where date > startDate and date < endDate)
//A list with alot of dates and numbers, most of the dates are the same. List contains roughly 1000 items, but can be more and less.
for(var date = startDate; date < endDate; date = date.AddDay(1){
listItem = objs.Where(x => x.Day = date).Sum(x => x.alotOfNUmbers);
}
Now since i don't care what day i calculate first, i thought i could do this.
Parallel.For(startDate, endDate, day => {
listItem = objs.Where(x => x.Day = date).Sum(x => x.alotOfNUmbers);
}
But how do i make it step dates ?

You can make a Range and iterate over it with Parallel.ForEach :
// not tested
var days = Enumerable
.Range(0, (endDate-startDate).Days) // check the rounding
.Select(i => startDate.AddDays(i));
Parallel.ForEach(days, day => ....)
Alternatively, you could use PLinq over the original source, probably faster. Roughly:
// not tested
var sums = objs.AsParallel().GroupBy(x => x.date).Select(g => g.Sum(i => i.number));

All the overloads of Parallel.For take two integer variables for start and end. I also don't see any version which would support something like a step so you can't just use the tick count of a DateTime as the loop variable.
But it should be easy to use a Parallel.ForEach instead, when you create an IEnumerable<DateTime> as the source sequence.
var source = Enumerable.Range(0, (endDate - startDate).Days)
.Select(t => startDate.AddDays(t));
Add +1 to the count parameter if the endDate is inclusive.

Ok after a few days search i figured if i placed all days in an array and "whiled" through it. It gives a pretty good result. With code easy to read
var start = new DateTime(2014, 09, 09);
var end = new DateTime(2014, 10, 01);
var listOfDays = new List<DateTime>();
int i = 0;
for (var day = start; day < end; day = day.AddDays(1))
{
listOfDays.Add(day);
}
Parallel.ForEach(listOfDays.ToArray(), currentDay =>
{
for (var d = new DateTime(currentDay.Year, currentDay.Month, currentDay.Day, 0, 0, 0); d < new DateTime(currentDay.Year, currentDay.Month, currentDay.Day, 23, 59, 59); d = d.AddSeconds(5))
{
var str = "Loop: " + i + ", Date: " + d.ToString();
Console.WriteLine(str);
}
i++;
});
Console.Read();