Summarize Data

[rbb]

I am trying to get the number of entries per hour in a log file.

 

The data looks like this:

12/5/04 8:15 AM - blah blah blah

...

12/5/04 8:50 AM - blah blah blah

...

12/5/04 10:59 AM - blah blah blah

 

 

The results should be something like

 

Date - Count

12/5/04 8 AM - 15

12/5/04 9 AM - 21

12/5/04 10 AM - 2

 

In VB6 I used to create a temporary ADO recordset to summarize data. I would filter by the date, and if a match was found, increment the counter. Otherwise, create a new record and set its counter to one.

 

The question is - what is the proper way to do this same summarizing in .Net? Arraylist, DataSet? And can you please provide an example?

 

Thanks,

 

Rob

[mskeel]

A hashtable could work well here. You would parse out the date and use that as the key, then increment an intiger you store in the bucket. Something like (this is rough, from the hip, but should give you a general idea.)

 

 

 

'while looping through, reading your log file
'parse out the date and time

if hashy.ContainsKey(dateAndTimeYouParsed) then
  hashy(dateAndTimeYouParsed) += 1
else 'the key is not in, so you must initilize
  hashy.Add(dateAndTimeYourParsed, 1)
end if

'read the next line
'end the loop

[neodammer]

You could also since that data isnt too congested use a file for datastorage.

[rbb]

You could also since that data isnt too congested use a file for datastorage.

 

I acutally ended up using a dataset. Thank you for the replies though.

 

Rob

 

Dim ds as DataSet
ds = New DataSet
Dim myColumn As DataColumn
Dim d(1) As DataColumn

ds.Tables.Add("tmp")
myColumn = New DataColumn
myColumn.ColumnName = "year-month"
myColumn.DataType = GetType(String)
myColumn.Unique = True
ds.Tables("tmp").Columns.Add(myColumn)
d(0) = myColumn

myColumn = New DataColumn

myColumn.ColumnName = "count"
myColumn.DataType = GetType(Integer)
myColumn.DefaultValue = 0
ds.Tables("tmp").Columns.Add(myColumn)

ds.Tables("tmp").PrimaryKey = d

'And later on...

With ds.Tables("tmp")
  dr = .Rows.Find(my_Date)
  If dr Is Nothing Then
     dr = .NewRow
     dr(0) = my_Date
     dr(1) = Integer.Parse(my_Counter)
     .Rows.Add(dr)
  Else
     dr(1) += Integer.Parse(Counter)
  End If
End With

[mskeel]

Looks like that'll work well.

 

It appears to be a lot of overhead, settign up and naming columns, declaring primary keys, etc. A hashtable would take far less code and accomplish the exact same thing...all for free. But, you do know the problem at hand and there may be intricacies I do not understand completely.

 

What does using the dataset buy you? I am not familiar with that data sctructure. Why did you decide to go with it?