60 minute series

[futurenets]

its 05:43 Eastern Time on my PC and I've run the attached study on the following setup.

http://www.futurenets.co.uk/eSignal/esig2.jpg

shouldn't the Chart data be in 30 min intervals?
shouldn't the bar closes match up?

[futurenets]

attached study

[futurenets]

my fault. closes are ok see new study. just the Chart data times?

[futurenets]

the time problem appears to be only with historical data. As new bars are created in real-time I think the times are ok.

[futurenets]

also problem appears to be with EUR A0-FX but not with GBP A0-FX

[futurenets]

GBP A0-FX appears to have missing bars in 60min series.

[futurenets]

ignore last post. GBP A0-FX now has same problem as EUR A0-FX. Issue appears to be intermittent.

[futurenets]

still have a problem.



From the image you can see the first 2 rows are ok …

Hour Hour
| |
-1 60min( 1.5737, 4, 0, ...), Chart( 1.5737, 4, 30, ...)
-2 60min( 1.5757, 3, 0, ...), Chart( 1.5734, 4, 0, ...)

The first number is the bar index (i.e. last bar is 0, one before that is -1 and so on).

If you follow the rows down you will see how 60min becomes out of sync with 30min.

This problem is also intermittent and may be a problem only with certain symbols or US vs UK location or times of the day (this was taken 5am ET).

[futurenets]

3 hours later, still out-of-sync ...

[futurenets]

please remove previous post (DONE)

A full efs time series can only be created if the chart interval is less than time series interval and the chart interval is a factor of the time series interval.

Otherwise a partial efs time series will be created dependant on the intervals used.

[futurenets]

final conclusion:


1) A full efs alternative time series can only be created if the chart interval is less than alternative time series interval and the chart interval is a factor of the alternative time series interval.

2) Otherwise an incomplete alternative time series will be created depending on the intervals used.

3) However the attached SeriesTest4 shows that around 2am ET (from the UK), null values are also created during load-time in the alternative time series and this problem resolves itself around 9amET.

[futurenets]

could someone rename this to something like "using efs time series" so others can find it