[R-SIG-Finance] Aggregating time series by key and time

[Robert A'gata]

Hi,

I have a time series. Mostly trade time series which contain time,
tradeType, tradePrice and tradeSize columns. Says

                                     trdType trdPrice trdSize
2011-10-07 12:17:50       3 1357.998       9
2011-10-07 12:29:10       1 1351.480       4
2011-10-07 12:30:56       3 1361.366       8
2011-10-07 12:30:56       3 1361.366       3
2011-10-07 12:32:33       3 1316.613       4
2011-10-07 12:34:20       2 1321.174       4
2011-10-07 12:35:17       3 1345.164      10
2011-10-07 12:40:23       1 1341.528      10
2011-10-07 12:40:23       2 1331.528       4
2011-10-07 12:43:03       2 1332.578       8

I'd like to aggregate trdSize for trades occurred at the same
timestamp and of the same trade type. So it is not only aggregating by
time. I'm wondering if there is a good way to do this. Any advice
would be greatly appreciated. For reference, the output of such a
function for the above example would be:

                                     trdType trdPrice trdSize
2011-10-07 12:17:50       3 1357.998       9
2011-10-07 12:29:10       1 1351.480       4
2011-10-07 12:30:56       3 1361.366      11
2011-10-07 12:32:33       3 1316.613       4
2011-10-07 12:34:20       2 1321.174       4
2011-10-07 12:35:17       3 1345.164      10
2011-10-07 12:40:23       1 1341.528      10
2011-10-07 12:40:23       2 1331.528       4
2011-10-07 12:43:03       2 1332.578       8

Because at 2011-10-07 12:30:56 we have 2 trades of trdType=3 occurred
with sizes 8 and 3. However, at 2011-10-07 12:40:23, we have 2 trades
occurred at the same price but they differ in trdType. So we do not
perform aggregation. Thank you.

Best regards,

Robert