Thinksamuel

I have a coll where the first column is a unix timestamp value and the second column a value (that I will use to trigger notes). To read through the coll, I have a general counter. When the counter value is equal to the unix timestamp, the coll will bang out the associated value. for example: if the counter hits 1448974144 and the timestamp value on the line in the coll is the same value, then the coll will output the associated value and set the next timestamp in place to be evaluated.

The problem arises when wanting to read multiple colls in parallel: I want to enable the user to select different calls to be read. He can choose from a menu which colls to read. When he loads another coll, the counter should be able to find the nearest timestamp and start trying t evaluate from there on. Quite complex but how should I tackle this in a logical way?

I have a coll where the first column is a unix timestamp value and the second column a value (that I will use to trigger notes). To read through the coll, I have a general counter. When the counter value is equal to the unix timestamp, the coll will bang out the associated value. for example: if the counter hits 1448974144 and the timestamp value on the  line in the coll is the same value, then the coll will output the associated value and set the next timestamp in place to be evaluated.

reading-multiple-calls-with-timestamps

The easiest way would be to be able to query entries based n that timestamp value instead of the index value. Any idea how that would be possible?

the timestamp values can be the indexes themselves. you could fill your coll like this : 1374192115,1374192115; 1374192116, 1374192116; 1374192117, 1374192117; etc... ( the second item could be anything i just repeated the timestamp value. )

the timestamp values can be the indexes themselves. you could fill your coll like this :  1374192115,1374192115; 1374192116, 1374192116; 1374192117, 1374192117; etc... ( the second item could be anything i just repeated the timestamp value. )


That would involve manipulating the datasets which would become a hassle when working with more than 100 files. Furthermore, if you are going to use the timestamp as an index you will end up with 86400 values for one day. Imagine that the dataset is tracking 10 days. That would make the file very big.

Furthermore it actually makes the problem worse: The clock in the main patch will send the timestamp value to te coll to be matched. So imagine that our clock is ticking, it will match every second so it will output a value second which I do not want.

That would involve manipulating the datasets which would become a hassle when working with more than 100 files. Furthermore, if you are going to use the timestamp as an index you will end up with 86400 values for one day. Imagine that the dataset is tracking 10 days. That would make the file very big.
Furthermore it actually makes the problem worse: The clock in the main patch will send the timestamp value to te coll to be matched. So imagine that our clock is ticking, it will match every second so it will output a value second which I do not want.


Reading multiple colls with timestamps