Databases
StackOverflow2013
Postsdbo
PO
 

Note that there are some explanatory texts on larger screens.

plurals
  1. PO
    primarykey
    data
    text
    <p>Consider a matrix as a set of rows. In each row there are indices of the smallest and largest valid values. If you know these indices, you can calculate number of valid values in that row in O(1). (by valid I mean a value between M[i,j] and M[i0,j0].)</p> <p>Now, the matrix is sorted. Lets take the lower bound: (i,j).</p> <ul> <li><p>If you want to find the index of smallest valid value in previous row, it has to lie on the right side of (i,j). This is because right above (i,j) there has to be invalid (too small) value.</p></li> <li><p>If you want to find the index of smallest valid value in next row, it has to lie on the left side of (i,j) (or right below it).</p></li> </ul> <p>So you need to walk at most 2n "steps" across the matrix to find lower bound indices of every row. The same is with upper bound. So your walking is O(n), then calculating number of valid values for each row is O(n), therefore total time is O(n).</p> <p>Using this algorithm it is possible to solve the median problem. Firstly note that if you calculate a solution to previous problem, you can pick a valid value at random in linear time using indices for boundary values in every row. Median can be then calculated by a bisection algorithm:</p> <pre><code>selection_find(M, i0,j0, i2,j2, K): # Find K-th smallest number between M(i0,j0) and M(i2,j2) # assumption: M(i0,j0)&lt;M(i2,j2) N := number of values between M(i0,j0) and M(i2,j2) # assumption: k&lt;N Pick at random i1,j1 so that M(i0,j0)&lt;M(i1,j1)&lt;M(i2,j2) L := number of values between M(i0,j0) and M(i1,j1) if L==K: The answer is M(i1,j1) if L&lt;K: The answer is selection_find(M, i1,j1, i2,j2, K-L) if L&gt;K: The answer is selection_find(M, i0,j0, i1,j1, K) median_find(M): The answer is selection_find(M, 1,1, n,n, n²/2) </code></pre> <p>Each step takes O(N). There will be O(log N²)=O(2log N)=O(log N) steps (each steps should halve number of values considered). Therefore the total complexity is O(NlogN).</p>
    singulars
    1. This table or related slice is empty.
    1. POCompletely Sorted matrix
      singulars
      1. PTQuestion
    1. PTAnswer
    1. USliori
    1. USliori
    plurals
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. POCompletely Sorted matrix
      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. VTAcceptedByOriginator
    1. COThanks a lot Liori !
      singulars
      1. PO
      1. USXero
    2. COWould you please check whether the line `if L>K: The answer is selection_find(M, i0,j0, i1,j1, K)` should be `if L>K: The answer is selection_find(M, i0,j0, i1,j1, L-K)`? A user with low rep comments about this.
      singulars
      1. PO
      1. USnhahtdh
    3. CO@nhahtdh: Oh, thank you for message, I would have missed it. Yes, I see Michel's answer, and I think he's mistaken, although I'm a little dizzy right now and I am not sure. The thing is, this case is when our random shot went too far: we know that (lower bound) < (median) < (our shoot) < (upper bound). Therefore we can reduce the “interval” in which we're looking for median to the part below our shot, and we're changing the upper bound. But the K does not need adjustment—it does only in the opposite case, where we're adjusting the lower bound. Although, as I say, I'm not at my best right now…
      singulars
      1. PO
      1. USliori
 

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