Databases
StackOverflow2013
Postsdbo
PO
 

Note that there are some explanatory texts on larger screens.

plurals
  1. PO
    primarykey
    data
    text
    <p>For the first function you're looking for the cumulative number of periods during which series <code>x</code> is lower/higher than <code>y</code>. For that you can use this handy function <code>CumCount()</code> built from <code>cummax</code>. First some sample data:</p> <pre><code>set.seed(1) x &lt;- sample(1:5,20,T) y &lt;- sample(1:5,20,T) CumCount &lt;- function(x) { z &lt;- cumsum(x) z - cummax(z*(!x)) } CumLow = CumCount(x&lt;y) CumHigh = CumCount(x&gt;y) </code></pre> <p>For your second computation, you're trying to find the <em>cumulative minimum</em> <code>x</code> value <em>within each period</em> during which <code>x &lt; y</code>. For this the <code>rle</code> function is very useful ("run-length-encoding").</p> <pre><code># runs equals the length of each phase (x &lt; y or x &gt; y) runs &lt;- rle(CumLow &gt; 0)$lengths # starts is the number of periods prior to each phase... starts &lt;- c(0,cumsum(runs)[-length(runs)]) #... which we use to build "blocks", a list of indices of each phase. blocks &lt;- mapply( function(x,y) x+y, starts, lapply(runs,seq)) # now apply the cummin function within each block: # (remember to mask it by CumLow &gt; 0 -- # we only want to do this within the x&lt;y phase) BlockCumMin &lt;- unlist(sapply(blocks, function(blk) cummin(x[blk]))) * (CumLow &gt; 0) </code></pre> <p>Now we put it all together:</p> <pre><code> &gt; cbind(x,y, CumLow, CumHigh, BlockCumMin) x y CumLow CumHigh BlockCumMin [1,] 3 4 1 0 3 [2,] 4 2 0 1 0 [3,] 2 2 0 0 0 [4,] 2 5 1 0 2 [5,] 4 4 0 0 0 [6,] 2 2 0 0 0 [7,] 4 1 0 1 0 [8,] 1 3 1 0 1 [9,] 2 5 2 0 1 [10,] 1 3 3 0 1 [11,] 2 5 4 0 1 [12,] 1 4 5 0 1 [13,] 4 2 0 1 0 [14,] 5 3 0 2 0 [15,] 4 1 0 3 0 [16,] 4 1 0 4 0 [17,] 3 4 1 0 3 [18,] 3 1 0 1 0 [19,] 5 3 0 2 0 [20,] 4 4 0 0 0 </code></pre> <p>Note that this problem is related to <a href="https://stackoverflow.com/questions/5012516/count-how-many-consecutive-values-are-true/5016069#5016069">this question</a></p> <p><strong><em>Update.</em></strong> For the more general case where you have a <code>series</code> vector, a <code>lengths</code> vector (of same length as <code>series</code>), and you want to produce a result called <code>BlockMins</code> where <code>BlockMins[i]</code> is the minimum of the <code>lengths[i]</code> block of <code>series</code> ending at position <code>i</code>, you could do the following. Since the lengths are arbitrary, this is no longer a cumulative min; for each <code>i</code> you have to take the min of the <code>length[i]</code> elements of <code>series</code> ending at position <code>i</code>:</p> <pre><code>set.seed(1) series &lt;- sample(1:5,20,T) lengths &lt;- sample(3:5,20,T) BlockMins &lt;- sapply(seq_along(lengths), function(i) min( series[ i : max(1, (i - lengths[i]+1)) ]) ) &gt; cbind(series, lengths, BlockMins) series lengths BlockMins [1,] 1 5 1 [2,] 1 4 1 [3,] 3 3 1 [4,] 4 4 1 [5,] 5 3 3 [6,] 1 4 1 [7,] 1 5 1 [8,] 4 3 1 [9,] 2 5 1 [10,] 2 4 1 [11,] 1 5 1 [12,] 2 5 1 [13,] 2 3 1 [14,] 2 4 1 [15,] 4 5 1 [16,] 3 5 2 [17,] 5 3 3 [18,] 1 4 1 [19,] 5 3 1 [20,] 3 3 1 </code></pre>
    singulars
    1. This table or related slice is empty.
    1. POIs it posible to optimize (vectorize) these two functions for better performance
      singulars
      1. PTQuestion
    1. PTAnswer
    1. USCommunity
    1. USPrasad Chalasani
    plurals
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. POIs it posible to optimize (vectorize) these two functions for better performance
      singulars
      1. PTQuestion
    1. This table or related slice is empty.
    1. VO
      singulars
      1. PO
      1. This table or related slice is empty.
      1. VTUpMod
    2. VO
      singulars
      1. PO
      1. This table or related slice is empty.
      1. VTUpMod
    3. VO
      singulars
      1. PO
      1. This table or related slice is empty.
      1. VTUpMod
    1. COPrasad, thank you very much. This is really insightful. Is cummin function really the same as runMin in TTR package?
      singulars
      1. PO
      1. USSamo
    2. COYou're welcome. The `runMin` function in `TTR` has more functionality than `cummin` -- it allows you to do the `min` over a moving window. In fact it calls `cummin` if the `cumulative` optional arg = `TRUE`. You can see how it's implemented by typing `runMin` at the R console -- remember R is open-source!
      singulars
      1. PO
      1. USPrasad Chalasani
    3. COSo I can simply exchange cummin with runMin from TTR in your code and everyting will work as in my nonperformant example above except it will be much faster? Beacuse basically I ned two separate functions as stated in my question with different moving window. Thnx.
      singulars
      1. PO
      1. USSamo
 

Querying!

 
Guidance
A row detail

Detail views are divided into sections. All the information in the data section comes from columns in the selected row. The other sections display data from other, related rows.

Related data can be related in a to-one or a to-many fashion. Captions of data related in a to-many fashion link to a list view showing a filtered view of the table.

Try moving around until you find a non-empty to-many entry and click on the label to get to one. You can move back to the root by clicking on the database name in the header.

SQuiL has stopped working due to an internal error.

If you are curious you may find further information in the browser console, which is accessible through the devtools (F12).

Reload