Rally - Interpret DragAndDropRank
Article ID: 47752
Updated On:
Products
Issue/Introduction
When using the API to inspect ranking, it is not a numeric value and difficult to determine what the actual rank of the work item is. This article gives an example of how to interpret the 64 character value of DragAndDropRank into a numeric value.
Please note, the method described below is for information purposes only and is NOT supported. It demonstrates a rough approximation of the method Rally developers used to make the ranks consistent after the old numeric Rank attribute was replaced with the new alpha-numeric DragAndDropRank attribute. There are no guarantees that when this method is followed as described a complete match to the results of Rally's internal ranking algorithm can be achieved.
Resolution
The rank values are base 94 values, using ASCII table but without the control characters (see the char[] digits array below). When we return WSAPI 1.x rank values we convert them to twelve digit decimals to simulate our old rank values. To do that we pull the 6 "most significant" characters from the left hand side and convert each value to long, much like hexadecimal conversion. The customer could use a larger number of significant digits, but we found you need several million artifacts in a workspace before you start hitting collisions in this space.
Let's say WSAPI 2.x rank is:
,~~t~Bj^iEn{qqS5!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
We would use first 6 "most significant" characters from the left hand side:
,~~t~B
and it converts to 88068394267
If you query the same User Story in 2.x and 1.x and compare the DragAndDropRank vs the Rank value you can double check your calculations.
Here are some java code snippets to give a sense of how it works. It won't compile because some tweaking was done to them to have it make sense in this context. This code serves as an illustration only.
The 6 character number above would become raw in the below method; essentially we calculate the numeric value of each letter then multiply it by the appropriate base.
public long convertToLong(String raw) {
long ret = 0L;
char[] chars = raw.toCharArray();
int power = 0;
for (int i = chars.length -1 ; i >=0; i--) {
long multiplier = (long) Math.pow(93, power);
long multiplicand = lookup.get(chars[i]);
ret +=(multiplicand * multiplier);
power++;
}
return ret;
}
protected static final char[] digits = {
'!', '"', '#', '$', '%', '&', '\'', '(', ')', '*', '+', ',', '-', '.', '/',
'0', '1', '2', '3', '4', '5', '6', '7', '8', '9',
':', ';', '<', '=', '>', '?', '@',
'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z',
'[', '\\', ']', '^', '_', '`',
'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', 'x', 'y', 'z',
'{','|','}','~'
};
private static Map<Character, Integer> lookup;
static {
lookup = newHashMap();
for (int i = 0; i < digits.length; i++) {
lookup.put(digits[i], i);
}
}
EXAMPLE:
Let's say PortfolioItem/Feature query returns three objects. The query's fetch parameter included FromattedID and DragAndDropRank attributes:
Results:
[
{
FormattedID: F33,
DragAndDropRank: "P!!$(~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~"
},
{
FormattedID: F34,
DragAndDropRank: "P!!$)~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~"
},
{
FormattedID: F35,
DragAndDropRank: "P!!$*~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~"
}
]
The method below is an approximation. The precision will vary depending on the number of characters interpreted: the larger the number the closer it comes to being accurate.
Here is the array of characters that are used in DragAndDropRank. It is similar to ASCII table, but without the control characters
?'!', '"', '#', '$', '%', '&', '\'', '(', ')', '*', '+', ',', '-', '.', '/',? '0', '1', '2', '3', '4', '5', '6', '7', '8', '9',? ':', ';', '<', '=', '>', '?', '@',? 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z',? '[', '\\', ']', '^', '_', '`',? 'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', 'x', 'y', 'z',? '{','|','}','~'
We will use base 94 and accumulate values from right to left. In this example we will start with the fifth character in the DragAndDropRank (fifth from the left) and go from right to left, but in many cases more precision is required.
This is an artificial example: only three features were created in the empty project, and we are going to interpret the initial DragAndDropRank values generated with the original revision of the artifacts.? Due to the contrived nature of this example? 5 characters is enough. Starting with the 6th character the values for rank are identical for all three features.
Here are the characters that happen to be used in the DragAndDropRank values in this exmaple:
'!' is at index 0
'$' is at index 3
'(' is at index 7
')' is at index 8
'*' is at index 9
'+' is at index 10
'P' is at index 47
Feature F35:
P!!$* resolves to 3669520403 based on these calculations:
character: *
(9 * (94^0=1)) = 9
accumulated = 9
character: $
(3 * (94^1=94)) = 282
accumulated = 291
character: !
(0 * (94^2=8836)) = 0
accumulated = 291
character: !
(0 * (94^3=830584)) = 0
accumulated = 291
character: P
(47 * (94^4=78074896)) = 3669520112
accumulated = 3669520403
Feature F34:
P!!$) resolves to 3669520402 based on these calculations:
character: )
(8 * (94^0=1)) = 8
accumulated = 8
character: $
(3 * (94^1=94)) = 282
accumulated = 290
character: !
(0 * (94^2=8836)) = 0
accumulated = 290
character: !
(0 * (94^3=830584)) = 0
accumulated = 290
character: P
(47 * (94^4=78074896)) = 3669520112
accumulated = 3669520402
Feature F33:
P!!$( resolves to 3669520401 based on these calculations:
character: (
(7 * (94^0=1)) = 7
accumulated = 7
character: $
(3 * (94^1=94)) = 282
accumulated = 289
character: !
(0 * (94^2=8836)) = 0
accumulated = 289
character: !
(0 * (94^3=830584)) = 0
accumulated = 289
character: P
(47 * (94^4=78074896)) = 3669520112
accumulated = 3669520401
If we stack the interpreted numeric values in ascending order they express the relative ranking, F35 being of highest rank and F33 being of lowest.
Feature F33: 3669520401
Feature F34: 3669520402
Feature F35: 3669520403
Additional Information
Feedback
