SortedMap: getting value for largest key less or equal a given

[Andreas Leitgeb]

From: "Andreas Leitgeb" <andreas.leitgeb@1:261/38.remove-yy0-this>

From: "Andreas Leitgeb" <andreas.leitgeb@1:261/38.remove-s5y-this>

From: Andreas Leitgeb <[email protected]>

I've got an approach like the following, but I'm not entirely happy with it.
(see embedded comments) Error checking left out only for brevity-of-post's
sake.

<sscce>
class StepFunction<K,V> {
SortedMap<K,V> m_map = new TreeMap<K,V>();
public void put(K k,V v) { m_map.put(k,v); } // for demo-fill

/** @return the value for the largest key in the map
that is less OR equal to the given parameter. */

public V value(K k) {
// not really correct for generic use. In my usecase, K is
// actually Long, so I just add one to k to make it work.
return m_map.get(m_map.headMap(k).lastKey());

// I'm a bit unhappy about headMap's "open end",
// and also about the lack of some method like
// lastEntry() or lastKeysValue() in SortedMap,
// requiring one to look up the lastKey in the map,
// although the map had "its finger on it" just before.
// Did I miss something simple and obvious?
}
// demo-helper
void checkVal(K k, V v) {
System.out.println( map.value(k) + " should be " + v);
}
public static void main(String[] args) {
StepFunction<Integer,Double> sf = new StepFunction<>()
sf.put(Integer.MIN_VALUE, -1.0);
sf.put(0, 0.0); sf.put(2, 1.0);

sf.checkVal(-1 , -1.0);
sf.checkVal( 0 , 0.0);
sf.checkVal( 1 , 0.0);
sf.checkVal( 2 , 1.0);
sf.checkVal(Integer.MAX_VALUE , 1.0);
}
}
</sscce>

PS: No need to offer "solutions" involving linear search.
I could have come up with one, myself, if I wanted one.

-+- BBBS/Li6 v4.10 Dada-1
+ Origin: Prism bbs (1:261/38)
-+- Synchronet 3.16a-Win32 NewsLink 1.98
Time Warp of the Future BBS - telnet://time.synchro.net:24

-+- BBBS/Li6 v4.10 Dada-1
+ Origin: Prism bbs (1:261/38)
-+- Synchronet 3.16a-Win32 NewsLink 1.98
Time Warp of the Future BBS - telnet://time.synchro.net:24

--- BBBS/Li6 v4.10 Dada-1
* Origin: Prism bbs (1:261/38)
--- Synchronet 3.16a-Win32 NewsLink 1.98
Time Warp of the Future BBS - telnet://time.synchro.net:24

 

[Andreas Leitgeb]

To: Andreas Leitgeb
From: "Andreas Leitgeb" <andreas.leitgeb@1:261/38.remove-yy0-this>

To: Andreas Leitgeb
From: "Andreas Leitgeb" <andreas.leitgeb@1:261/38.remove-s5y-this>

To: Andreas Leitgeb
From: Andreas Leitgeb <[email protected]>

I've got an approach like the following, but I'm not entirely
happy with it...

It took me about an hour to compose that previous post, and not only did a bug
still hide in it ("map" instead of "m_map"), but also: only five minutes after
posting, it occurred to me that I could also look at TreeMap's methods, rather
than only at SortedMap's, and thereby stumbled over the (new in 1.6) interface
NavigableMap.

PS: return m_map.floorEntry(k).getValue(); // Case closed.

-+- BBBS/Li6 v4.10 Dada-1
+ Origin: Prism bbs (1:261/38)
-+- Synchronet 3.16a-Win32 NewsLink 1.98
Time Warp of the Future BBS - telnet://time.synchro.net:24

-+- BBBS/Li6 v4.10 Dada-1
+ Origin: Prism bbs (1:261/38)
-+- Synchronet 3.16a-Win32 NewsLink 1.98
Time Warp of the Future BBS - telnet://time.synchro.net:24

--- BBBS/Li6 v4.10 Dada-1
* Origin: Prism bbs (1:261/38)
--- Synchronet 3.16a-Win32 NewsLink 1.98
Time Warp of the Future BBS - telnet://time.synchro.net:24

 

[markspace]

To: Andreas Leitgeb
From: "markspace" <markspace@1:261/38.remove-yy0-this>

To: Andreas Leitgeb
From: "markspace" <markspace@1:261/38.remove-s5y-this>

To: Andreas Leitgeb
From: markspace <-@.>

It took me about an hour to compose that previous post, and not
only did a bug still hide in it ("map" instead of "m_map"), but
also: only five minutes after posting, it occurred to me that I
could also look at TreeMap's methods, rather than only at
SortedMap's, and thereby stumbled over the (new in 1.6) interface
NavigableMap.

PS: return m_map.floorEntry(k).getValue(); //
Case closed.

It is my contention that any requests for help should contain a careful
explanation of the problem and attempted solutions, plus an SSCCE. Then the
author should save the request on disc and go to lunch. If no solution was
discovered during lunch, then the request should be sent.

I think there's a Dilbert cartoon about this.

-+- BBBS/Li6 v4.10 Dada-1
+ Origin: Prism bbs (1:261/38)
-+- Synchronet 3.16a-Win32 NewsLink 1.98
Time Warp of the Future BBS - telnet://time.synchro.net:24

-+- BBBS/Li6 v4.10 Dada-1
+ Origin: Prism bbs (1:261/38)
-+- Synchronet 3.16a-Win32 NewsLink 1.98
Time Warp of the Future BBS - telnet://time.synchro.net:24

--- BBBS/Li6 v4.10 Dada-1
* Origin: Prism bbs (1:261/38)
--- Synchronet 3.16a-Win32 NewsLink 1.98
Time Warp of the Future BBS - telnet://time.synchro.net:24