Databases
StackOverflow2013
Postsdbo
PO
 

Note that there are some explanatory texts on larger screens.

plurals
  1. PO
    primarykey
    data
    text
    <p>For each combination of components to be totally new for each distinct Player, it is necessary and sufficient that all components occur for the first time for that Player. Therefore we keep track of first occurrences of each component and drop any observations with later occurrences of that component. </p> <p>The same toy data (but with some shorter variable names): </p> <pre><code>input Product Player_id C1 C2 C3 1 1 1 2 3 2 1 1 5 9 3 1 7 8 9 4 2 4 5 6 5 2 4 5 9 6 2 7 8 9 end </code></pre> <p>It is not clear whether <code>Product</code> plays the same role as observation number: regardless of that we can create a variable holding observation numbers. </p> <pre><code>gen long obs = _n </code></pre> <p>Now we are going to loop over components, 9 of them in this example. </p> <pre><code>qui forval i = 1/9 { </code></pre> <p>For each component: it's found if it's any of the three components. That is an indicator variable, 1 if true and 0 if false: </p> <pre><code> gen found = C1 == `i' | C2 == `i' | C3 == `i' </code></pre> <p>There is a neat alternative: </p> <pre><code> gen found = inlist(`i', C1, C2, C3) </code></pre> <p>We want to keep track of the first occurrence, which is the minimum observation number if <code>found</code> is true, and we do that separately by <code>Player_id</code>. The division here by the indicator variable produces <code>obs</code> if <code>found</code> is 1 and missing otherwise. The missings are just ignored by <code>egen</code> unless all values are returned as missing. </p> <pre><code> egen first = min(obs / found), by(Player_id) </code></pre> <p>Now what we do is <code>drop</code> occurrences if they are not the first, and <code>drop</code> our accounting variables, so they can be re-created next time round the loop: </p> <pre><code> drop if obs &gt; first &amp; found drop first found } </code></pre> <p>Here's the result: </p> <pre><code> +-----------------------------------------+ | Product Player~d C1 C2 C3 obs | |-----------------------------------------| 1. | 1 1 1 2 3 1 | 2. | 3 1 7 8 9 3 | 3. | 4 2 4 5 6 4 | 4. | 6 2 7 8 9 6 | +-----------------------------------------+ </code></pre> <p>Note that this algorithm is sensitive to the order of the observations. If you start with <code>1 5 9</code> then neither <code>1 2 3</code> or <code>7 8 9</code> are acceptable in addition. Similarly with <code>4 5 9</code>; <code>4 5 6</code> and <code>7 8 9</code> don't make the cut. </p>
    singulars
    1. This table or related slice is empty.
    1. POStata: Identifying unique observations that differ on all variables, by group
      singulars
      1. PTQuestion
    1. PTAnswer
    1. USNick Cox
    1. USNick Cox
    plurals
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. POStata: Identifying unique observations that differ on all variables, by group
      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. VTAcceptedByOriginator
    1. COThanks a lot for your consideration, Nick. I have tested your solution but it appears that the order in which observations are eliminated matters for this problem. I have tried to develop an alternative solution and to sequentially eliminate products based on pairwise value comparisons within a group: http://stackoverflow.com/questions/18393424/stata-sequential-elimination-of-observations-based-on-pairwise-comparisons-with
      singulars
      1. PO
      1. USuser2700264
 

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