CloudTurbine Forums

Welcome to the CloudTurbine official Forums! Please join us to share your experiences, ask questions, share your ideas, and support the community.

Forum Replies Created

Viewing 4 posts - 1 through 4 (of 4 total)
  • Author
    Posts
  • in reply to: Suggestions for WebScan #7222
    Matt Miller
    Matt Miller
    Keymaster

    I have updated WebScan so you can set the number of vertical grids per plot. It is part of WebScan V3.1 release and will automatically update using CTweb. Other suggestions are on the to-do list.

    in reply to: Suggestions for WebScan #7220
    Matt Miller
    Matt Miller
    Keymaster

    I will make note of requests for features such as this. Your feedback helps prioritize future development. Thanks!

    in reply to: CTweb interface URL munge question #4632
    Matt Miller
    Matt Miller
    Keymaster

    Good catch. As of today’s CTweb.jar update, you can now fetch time, data, or both with the following munges:

    ?f=b (default) – both time and data
    ?f=t – time only
    ?f=d – data values only

    The prior version was meant to handle the data-only (f=d) case, but needed a bug fix to enable it.

    in reply to: CT segment+block folder structure #3404
    Matt Miller
    Matt Miller
    Keymaster

    I think you may be misinterpreting your folder structure. You state you have:
    BaseTime/Segment/Block
    I think what you really have is:
    BaseTime/(Segment)/Block/Point
    In your case the optional (Segment) is missing (the current default for CT API generated folders).
    Thus what you are calling a segment is really a block, and what you call a block is a point.
    See the online “CT File Structure” document for a big-picture explanation of all this.

    You normally would get zero or more segments, one or more blocks per segment, and always at least one point. The first segment and block are typically “0”, as in the leading digits of your odometer on a new car. The CT Java API automatically generates a placeholder folder hierarchy at startup to be in place for the general situation.

    In your case I think you have a single “Point” folder with “0” relative time per block. Normally you would have multiple points per block with point-times something like 0,10,20,… Or a single “packed” multi-point with a non-zero “duration”, e.g. “9000”. If you were to simply generate more points per block (or pack more points per channel), this format would become more clear.

    Whereas it is a legal CT file structure to collapse out “0” relative timestamp folders, once you have following non-zero folders you would end up with data channels at varying levels in the hierarchy, and arguably less consistent. For the case where you only have one segment, block, or point, collapsing “0” folders would make more sense, but it is not generally known how much data you may eventually generate.

    Try setting the ctw.autoflush(timePerBlock,blocksPerSegment) parameters. You can manually flush and/or disable segments by setting either parameter to zero. See the CTexample/CTblocktest.java for an example. If still confused please send a relevant code snippet.

    Final note: normally all this file structure stuff should be “magic” and not something to worry about (so long as it works!).

Viewing 4 posts - 1 through 4 (of 4 total)