Survey Database rounding

[sturner7277]

Is there an option to control the rounding in the survey database in C3D 2009? i have dug around and have not uncoverd anything. anyone have any iseas.

thanks,

Sean

[mathewkol]

Right click your database and select "Edit survey database settings".  Change the "Precision" of the desired property.

[sturner7277]

yea that changes the percision not how the program rouns the numbers up/down

[mathewkol]

I was assuming you meant that.  Can you explain further?

[sturner7277]

the options in the ambent settings is rounding to round up, round down, or truncate. i want to be able to control how c3d rounds the elevation in the survey point database.

[taren]

R U a -forgive me- a RE-RE?? The database is just it a database it holds the RAW (uncut) data that you give it. Once the data is in you manupulate it. A round is a round,   < 5 down, > 5 up. Is this some wried concept on yor planet??

[dewarp1]

Sean,

Rounding in Civil 3D is a bit odd.  "Round Normal" rounds 5s up OR down, to make the final rounded number even.  So 1.235 rounds to 1.24; and 1.245 also rounds to 1.24. 

This is discussed further in this post and also in the help file here.

I have a little workaround to get levels rounding and displaying so that all the 5's round up.  I have made an expression in my label style settings to add 0.000000001 to the level and use that expression value in my point labels.  This will always round 5's up as I would expect to happen.

Cheers

Phil Dewar
Fox and Associates

[sturner7277]

thanks Phil,

i was just hoping that i could control how the numbers show in the vista. i will push the percision out to three places for the elevation. so in the survey database vista it will display the way i want and the surveyers will see the numbers as they expect.

-Sean

[Sinc]

---

@dewarp1 wrote:

This will always round 5's up as I would expect to happen.

---

There is actually a very good argument for why the "round even" bias is much better.  It has to do with statistics and statistical averages...  The upshot is that if you always round all .5's up, then you get an overal bias toward higher numbers.  By using the "round even" bias, you correct for this, as some of the .5's round up, and some of the .5's round down.  It helps correct for the "round up" bias.

Strictly speaking, C3D's "round even" bias is better.  However, this difference is usally a very small percentage of the total, so often there is no practical difference between the two.  And of course, I understand the "practical difference" we see in the real world, e.g. review agencies that are overly-analytical and get all bent out of shape over rounding differences.  So I can understand the desire to control this behavior, so you can get "what the reviewer wants and expects" as opposed to "what is better".

Oh wait... Your're not worried about the appearance in labels, but in the survey database...?  Why, exactly?  I know it's often beyond the ken of reviewers in various agencies, but hopefully your in-house staff is not being confused by "rounding error"...?