Related
The Problem
On a server, I host ids in a json file. From clients, I need to mandate the server to intersect and sometimes negate these ids (the ids never travel to the client even though the client instructs the server its operations to perform).
I typically have 1000's of ids, often have 100,000's of ids, and have a maximum of 56,000,000 of them, where each value is unique and between -100,000,000 and +100,000,000.
These ids files are stable and do not change (so it is possible to generate a different representation for it that is better adapted for the calculations if needed).
Sample ids
Largest file sizes
I need an algorithm that will intersect ids in the sub-second range for most cases. What would you suggest? I code in java, but do not limit myself to java for the resolution of this problem (I could use JNI to bridge to native language).
Potential solutions to consider
Although you could not limit yourselves to the following list of broad considerations for solutions, here is a list of what I internally debated to resolve the situation.
Neural-Network pre-qualifier: Train a neural-network for each ids list that accepts another list of ids to score its intersection potential (0 means definitely no intersection, 1 means definitely there is an intersection). Since neural networks are good and efficient at pattern recognition, I am thinking of pre-qualifying a more time-consuming algorithm behind it.
Assembly-language: On a Linux server, code an assembly module that does such algorithm. I know that assembly is a mess to maintain and code, but sometimes one need the speed of an highly optimized algorithm without the overhead of a higher-level compiler. Maybe this use-case is simple enough to benefit from an assembly language routine to be executed directly on the Linux server (and then I'd always pay attention to stick with the same processor to avoid having to re-write this too often)? Or, alternately, maybe C would be close enough to assembly to produce clean and optimized assembly code without the overhead to maintain assembly code.
Images and GPU: GPU and image processing could be used and instead of comparing ids, I could BITAND images. That is, I create a B&W image of each ids list. Since each id have unique values between -100,000,000 and +100,000,000 (where a maximum of 56,000,000 of them are used), the image would be mostly black, but the pixel would become white if the corresponding id is set. Then, instead of keeping the list of ids, I'd keep the images, and do a BITAND operation on both images to intersect them. This may be fast indeed, but then to translate the resulting image back to ids may be the bottleneck. Also, each image could be significantly large (maybe too large for this to be a viable solution). An estimate of a 200,000,000 bits sequence is 23MB each, just loading this in memory is quite demanding.
String-matching algorithms: String comparisons have many adapted algorithms that are typically extremely efficient at their task. Create a binary file for each ids set. Each id would be 4 bytes long. The corresponding binary file would have each and every id sequenced as their 4 bytes equivalent into it. The algorithm could then be to process the smallest file to match each 4 bytes sequence as a string into the other file.
Am I missing anything? Any other potential solution? Could any of these approaches be worth diving into them?
I did not yet try anything as I want to secure a strategy before I invest what I believe will be a significant amount of time into this.
EDIT #1:
Could the solution be a map of hashes for each sector in the list? If the information is structured in such a way that each id resides within its corresponding hash key, then, the smaller of the ids set could be sequentially ran and matching the id into the larger ids set first would require hashing the value to match, and then sequentially matching of the corresponding ids into that key match?
This should make the algorithm an O(n) time based one, and since I'd pick the smallest ids set to be the sequentially ran one, n is small. Does that make sense? Is that the solution?
Something like this (where the H entry is the hash):
{
"H780" : [ 45902780, 46062780, -42912780, -19812780, 25323780, 40572780, -30131780, 60266780, -26203780, 46152780, 67216780, 71666780, -67146780, 46162780, 67226780, 67781780, -47021780, 46122780, 19973780, 22113780, 67876780, 42692780, -18473780, 30993780, 67711780, 67791780, -44036780, -45904780, -42142780, 18703780, 60276780, 46182780, 63600780, 63680780, -70486780, -68290780, -18493780, -68210780, 67731780, 46092780, 63450780, 30074780, 24772780, -26483780, 68371780, -18483780, 18723780, -29834780, 46202780, 67821780, 29594780, 46082780, 44632780, -68406780, -68310780, -44056780, 67751780, 45912780, 40842780, 44642780, 18743780, -68220780, -44066780, 46142780, -26193780, 67681780, 46222780, 67761780 ],
"H782" : [ 27343782, 67456782, 18693782, 43322782, -37832782, 46152782, 19113782, -68411782, 18763782, 67466782, -68400782, -68320782, 34031782, 45056782, -26713782, -61776782, 67791782, 44176782, -44096782, 34041782, -39324782, -21873782, 67961782, 18703782, 44186782, -31143782, 67721782, -68340782, 36103782, 19143782, 19223782, 31711782, 66350782, 43362782, 18733782, -29233782, 67811782, -44076782, -19623782, -68290782, 31721782, 19233782, 65726782, 27313782, 43352782, -68280782, 67346782, -44086782, 67741782, -19203782, -19363782, 29583782, 67911782, 67751782, 26663782, -67910782, 19213782, 45992782, -17201782, 43372782, -19992782, -44066782, 46142782, 29993782 ],
"H540" : [...
You can convert each file (list of ids) into a bit-array of length 200_000_001, where bit at index j is set if the list contains value j-100_000_000. It is possible, because the range of id values is fixed and small.
Then you can simply use bitwise and and not operations to intersect and negate lists of ids. Depending on the language and libraries used, it would require operating element-wise: iterating over arrays and applying corresponding operations to each index.
Finally, you should measure your performance and decide whether you need to do some optimizations, such as parallelizing operations (you can work on different parts of arrays on different processors), preloading some of arrays (or all of them) into memory, using GPU, etc.
First, the bitmap approach will produce the required performance, at a huge overhead in memory. You'll need to benchmark it, but I'd expect times of maybe 0.2 seconds, with that almost entirely dominated by the cost of loading data from disk, and then reading the result.
However there is another approach that is worth considering. It will use less memory most of the time. For most of the files that you state, it will perform well.
First let's use Cap'n Proto for a file format. The type can be something like this:
struct Ids {
is_negated #0 :Bool;
ids #1 :List(Int32);
}
The key is that ids are always kept sorted. So list operations are a question of running through them in parallel. And now:
Applying not is just flipping is_negated.
If neither is negated, it is a question of finding IDs in both lists.
If the first is not negated and the second is, you just want to find IDs in the first that are not in the second.
If the first is negated and the second is not, you just want to find IDs in the second that are not in the first.
If both are negated, you just want to find all ids in either list.
If your list has 100k entries, then the file will be about 400k. A not requires copying 400k of data (very fast). And intersecting with another list of the same size involves 200k comparisons. Integer comparisons complete in a clock cycle, and branch mispredictions take something like 10-20 clock cycles. So you should be able to do this operation in the 0-2 millisecond range.
Your worst case 56,000,000 file will take over 200 MB and intersecting 2 of them can take around 200 million operations. This is in the 0-2 second range.
For the 56 million file and a 10k file, your time is almost all spent on numbers in the 56 million file and not in the 10k one. You can speed that up by adding a "galloping" mode where you do a binary search forward in the larger file looking for the next matching number and picking most of them. Do be warned that this code tends to be tricky and involves lots of mispredictions. You'll have to benchmark it to find out how big a size difference is needed.
In general this approach will lose for your very biggest files. But it will be a huge win for most of the sizes of file that you've talked about.
Recently I ran into two sampling algorithms: Algorithm S and Algorithm Z.
Suppose we want to sample n items from a data set. Let N be the size of the data set.
When N is known, we can use Algorithm S
When N is unknown, we can use Algorithm Z (optimized atop Algorithm R)
Performance of the two algorithms:
Algorithm S
Time complexity: average number of scanned items is n(N+1)/n+1 (I compute the result, Knuth's book left this as exercises), we can say it O(N)
Space complexity: O(1) or O(n)(if returning an array)
Algorithm Z (I search the web, find the paper https://www.cs.umd.edu/~samir/498/vitter.pdf)
Time complexity: O(n(1+log(N/n))
Space complexity: in TAOCP vol2 3.4.2, it mentions Algorithm R's space complexity is O(n(1+log(N/n))), so I suppose Algorithm Z might be the same
My question
The model for Algorithm Z is: keep calling next method on the data set until we reach the end. So for the problem that N is known, we can still use Algorithm Z.
Based on the above performance comparison, Algorithm Z has better time complexity than Algorithm S, and worse space complexity.
If space is not a problem, should we use Algorithm Z even when N is known?
Is my understanding correct? Thanks!
Is the Postgres code mentioned in your comment actually used in production? In my opinion, it really should be reviewed by someone who has at least some understanding of the problem domain. The problem with random sampling algorithms, and random algorithms in general, is that it is very hard to diagnose biased sampling bugs. Most samples "look random" if you don't look too hard, and biased sampling is only obvious when you do a biased sample of a biased dataset. Or when your biased sample results in a prediction which is catastrophically divergent from reality, which will eventually happen but maybe not when you're doing the code review.
Anyway, by way of trying to answer the questions, both the one actually in the text of this post and the ones added or implied in the comment stream:
Properly implemented, Vitter's algorithm Z is much faster than Knuth's algorithm S. If you have a use case in which reservoir sampling is indicated, then you should probably use Vitter, subject to the code testing advice above: Vitter's algorithm is more complicated and it might not be obvious how to validate the implementation.
I noticed in the Postgres code that it just uses the threshold value of 22 to decide whether to use the more complicated code, based on testing done almost 40 years ago on hardware which you'd be hard pressed to find today. It's possible that 22 is not a bad threshold, but it's just a number pulled out of thin air. At least some attempt should be made to verify or, more likely, correct it.
Forty years ago, when those algorithms were developed, large datasets were typically stored on magnetic tape. Magnetic tape is still used today, but applications have changed; I think that you're not likely to find a Postgres installation in which a live database is stored on tape. This matters because the way you get data off a tape drive is radically different from the way you get data from a file server. Or a sharded distributed collection of file servers, which also has its particular needs.
Data on a reel of tape can only be accessed linearly, although it is possible to skip tape somewhat faster than you can read it. On a file server, data is random access; there may be a slight penalty for jumping around in a file, but there might not. (On the sharded distributed model, it might well be faster then linear reads.) But trying to read out of order on a tape drive might turn an input operation which takes an hour into an operation which takes a week. So it's very important to access the sample in order. Moreover, you really don't want to have to read the tape twice, which would take twice as long.
One of the other assumptions that was made in those algorithms is that you might not have enough memory to store the entire sample; in 1985, main memory was horribly expensive and databases were already quite large. So a common way to collect a large sample from a huge database was to copy the sampled blocks onto secondary memory, such as another tape drive. But there's a bit of a catch with reservoir sampling: as the sampling algorithm proceeds, some items which were initially inserted in the sample are later replaced with other items. But you can't replace data written on tape, so you need to just keep on appending the newly selected samples. What you do hold in random access memory is a list of locations of the sample; once you've finished selecting the sample, you can sort this list of locations and then use it to read out the final selection in storage order, skipping over the rejected items. That means that the temporary sample storage ends up holding both the final sample, and some number of later rejected items. The O(n(1+log(N/n))) space complexity in Algorithm R refers to precisely this storage, and it's actually a reasonably small multiplier, considering.
All that is irrelevant if you can just allocate enough random access storage somewhere to hold the entire sample. Or, even better, if you can directly read a data from the database. There could well still be good reasons to read the sample into local storage, but nothing stops you from updating a block of local storage with a different block.
On the other hand, in many common cases, you don't need to read the data in order to sample it. You can just take a list of items numbers, select a sample from that list of the desired size, and then set about acquiring the sample from the list of selected item numbers. And that presents a rather different problem: how to choose an unbiased sample of size k from a set of K item indexes.
There's a fast and simple solution to that (also described by Knuth, unsurprisingly): make an array of all the item numbers (say, the integers from 0 to K, and then shuffle the array using the standard Knuth/Fisher-Yates shuffle, with a slight modification: you run the algorithm from front to back (instead of back to front, as it is often presented), and stop after k iterations. At that point the first k elements in the partially shuffled array are an unbiased sample. (In fact, you don't need the entire vector of K indices, as long as k is much smaller than K. You're only going to touch O(k) of the values, and you can keep the ones you touched in a hash table of size O(k).)
And there's an even simpler algorithm, again for the case where the sample is small relative to the dataset: just keep one bit for each item in the dataset, which indicates that the item has been selected. Now select k items at random, marking the bit vector as you go; if the relevant bit is already marked, then that item is already in the sample; you just ignore that selection and continue with the next random choice. The expected number of ignored sample is very small unless the sample size is a significant fraction of the dataset size.
There's one other criterion which weighed on the minds of Vitter and Knuth: you'll normally want to do something with the selected sample. And given the amount of time it takes to read through a tape, you want to be able to start processing each item immediately as it is accepted. That precludes algorithms which include, for example, "sort the selected indices and then read the indicated items. (See above.) For immediate processing to be possible, you must not depend on being able to "deselect" already selected items.
Fortunately, both the quick algorithms mentioned at the end of point 2 do satisfy this requirement. In both cases, an item once selected will never be later rejected.
There is at least one use case for reservoir sampling which is still very much relevant: sampling a datastream which is too voluminous or too high-bandwidth to store. That might be some kind of massive social media feed, or it might be telemetry data from a large sensor array, or whatever. In that case, you might want to reduce the size of the datastream by extracting only a small sample, and reservoir sampling is a good candidate. However, that has nothing to do with the Postgres example.
In summary:
Yes, you can (and probably should) use Vitter's Algorithm Z in preference to Knuth's Algorithm S, even if you know how big the data set it.
But there are certainly better algorithms, some of which are outlined above.
Given two files containing list of words(around million), We need to find out the words that are in common.
Use Some efficient algorithm, also not enough memory availble(1 million, certainly not).. Some basic C Programming code, if possible, would help.
The files are not sorted.. We can use some sort of algorithm... Please support it with basic code...
Sorting the external file...... with minimum memory available,, how can it be implement with C programming.
Anybody game for external sorting of a file... Please share some code for this.
Yet another approach.
General. first, notice that doing this sequentially takes O(N^2). With N=1,000,000, this is a LOT. Sorting each list would take O(N*log(N)); then you can find the intersection in one pass by merging the files (see below). So the total is O(2N*log(N) + 2N) = O(N*log(N)).
Sorting a file. Now let's address the fact that working with files is much slower than with memory, especially when sorting where you need to move things around. One way to solve this is - decide the size of the chunk that can be loaded into memory. Load the file one chunk at a time, sort it efficiently and save into a separate temporary file. The sorted chunks can be merged (again, see below) into one sorted file in one pass.
Merging. When you have 2 sorted lists (files or not), you can merge them into one sorted list easily in one pass: have 2 "pointers", initially pointing to the first entry in each list. In each step, compare the values the pointers point to. Move the smaller value to the merged list (the one you are constructing) and advance its pointer.
You can modify the merge algorithm easily to make it find the intersection - if pointed values are equal move it to the results (consider how do you want to deal with duplicates).
For merging more than 2 lists (as in sorting the file above) you can generalize the algorithm for using k pointers.
If you had enough memory to read the first file completely into RAM, I would suggest reading it into a dictionary (word -> index of that word ), loop over the words of the second file and test if the word is contained in that dictionary. Memory for a million words is not much today.
If you have not enough memory, split the first file into chunks that fit into memory and do as I said above for each of that chunk. For example, fill the dictionary with the first 100.000 words, find every common word for that, then read the file a second time extracting word 100.001 up to 200.000, find the common words for that part, and so on.
And now the hard part: you need a dictionary structure, and you said "basic C". When you are willing to use "basic C++", there is the hash_map data structure provided as an extension to the standard library by common compiler vendors. In basic C, you should also try to use a ready-made library for that, read this SO post to find a link to a free library which seems to support that.
Your problem is: Given two sets of items, find the intersaction (items common to both), while staying within the constraints of inadequate RAM (less than the size of any set).
Since finding an intersaction requires comparing/searching each item in another set, you must have enough RAM to store at least one of the sets (the smaller one) to have an efficient algorithm.
Assume that you know for a fact that the intersaction is much smaller than both sets and fits completely inside available memory -- otherwise you'll have to do further work in flushing the results to disk.
If you are working under memory constraints, partition the larger set into parts that fit inside 1/3 of the available memory. Then partition the smaller set into parts the fit the second 1/3. The remaining 1/3 memory is used to store the results.
Optimize by finding the max and min of the partition for the larger set. This is the set that you are comparing from. Then when loading the corresponding partition of the smaller set, skip all items outside the min-max range.
First find the intersaction of both partitions through a double-loop, storing common items to the results set and removing them from the original sets to save on comparisons further down the loop.
Then replace the partition in the smaller set with the second partition (skipping items outside the min-max). Repeat. Notice that the partition in the larger set is reduced -- with common items already removed.
After running through the entire smaller set, repeat with the next partition of the larger set.
Now, if you do not need to preserve the two original sets (e.g. you can overwrite both files), then you can further optimize by removing common items from disk as well. This way, those items no longer need to be compared in further partitions. You then partition the sets by skipping over removed ones.
I would give prefix trees (aka tries) a shot.
My initial approach would be to determine a maximum depth for the trie that would fit nicely within my RAM limits. Pick an arbitrary depth (say 3, you can tweak it later) and construct a trie up to that depth, for the smaller file. Each leaf would be a list of "file pointers" to words that start with the prefix encoded by the path you followed to reach the leaf. These "file pointers" would keep an offset into the file and the word length.
Then process the second file by reading each word from it and trying to find it in the first file using the trie you constructed. It would allow you to fail faster on words that don't match. The deeper your trie, the faster you can fail, but the more memory you would consume.
Of course, like Stephen Chung said, you still need RAM to store enough information to describe at least one of the files, if you really need an efficient algorithm. If you don't have enough memory -- and you probably don't, because I estimate my approach would require approximately the same amount of memory you would need to load a file whose words were 14-22 characters long -- then you have to process even the first file by parts. In that case, I would actually recommend using the trie for the larger file, not the smaller. Just partition it in parts that are no bigger than the smaller file (or no bigger than your RAM constraints allow, really) and do the whole process I described for each part.
Despite the length, this is sort of off the top of my head. I might be horribly wrong in some details, but this is how I would initially approach the problem and then see where it would take me.
If you're looking for memory efficiency with this sort of thing you'll be hard pushed to get time efficiency. My example will be written in python, but should be relatively easy to implement in any language.
with open(file1) as file_1:
current_word_1 = read_to_delim(file_1, delim)
while current_word_1:
with open(file2) as file_2:
current_word_2 = read_to_delim(file_2, delim)
while current_word_2:
if current_word_2 == current_word_1:
print current_word_2
current_word_2 = read_to_delim(file_2, delim)
current_word_1 = read_to_delim(file_1, delim)
I leave read_to_delim to you, but this is the extreme case that is memory-optimal but time-least-optimal.
depending on your application of course you could load the two files in a database, perform a left outer join, and discard the rows for which one of the two columns is null
I am a graduate student of physics and I am working on writing some code to sort several hundred gigabytes of data and return slices of that data when I ask for it. Here is the trick, I know of no good method for sorting and searching data of this kind.
My data essentially consists of a large number of sets of numbers. These sets can contain anywhere from 1 to n numbers within them (though in 99.9% of the sets, n is less than 15) and there are approximately 1.5 ~ 2 billion of these sets (unfortunately this size precludes a brute force search).
I need to be able to specify a set with k elements and have every set with k+1 elements or more that contains the specified subset returned to me.
Simple Example:
Suppose I have the following sets for my data:
(1,2,3)
(1,2,3,4,5)
(4,5,6,7)
(1,3,8,9)
(5,8,11)
If I were to give the request (1,3) I would have the sets: (1,2,3),
(1,2,3,4,5), and (1,3,8,9).
The request (11) would return the set: (5,8,11).
The request (1,2,3) would return the sets: (1,2,3) and (1,2,3,4,5)
The request (50) would return no sets:
By now the pattern should be clear. The major difference between this example and my data is that the sets withn my data are larger, the numbers used for each element of the sets run from 0 to 16383 (14 bits), and there are many many many more sets.
If it matters I am writing this program in C++ though I also know java, c, some assembly, some fortran, and some perl.
Does anyone have any clues as to how to pull this off?
edit:
To answer a couple questions and add a few points:
1.) The data does not change. It was all taken in one long set of runs (each broken into 2 gig files).
2.) As for storage space. The raw data takes up approximately 250 gigabytes. I estimate that after processing and stripping off a lot of extraneous metadata that I am not interested in I could knock that down to anywhere from 36 to 48 gigabytes depending on how much metadata I decide to keep (without indices). Additionally if in my initial processing of the data I encounter enough sets that are the same I might be able to comress the data yet further by adding counters for repeat events rather than simply repeating the events over and over again.
3.) Each number within a processed set actually contains at LEAST two numbers 14 bits for the data itself (detected energy) and 7 bits for metadata (detector number). So I will need at LEAST three bytes per number.
4.) My "though in 99.9% of the sets, n is less than 15" comment was misleading. In a preliminary glance through some of the chunks of the data I find that I have sets that contain as many as 22 numbers but the median is 5 numbers per set and the average is 6 numbers per set.
5.) While I like the idea of building an index of pointers into files I am a bit leery because for requests involving more than one number I am left with the semi slow task (at least I think it is slow) of finding the set of all pointers common to the lists, ie finding the greatest common subset for a given number of sets.
6.) In terms of resources available to me, I can muster approximately 300 gigs of space after I have the raw data on the system (The remainder of my quota on that system). The system is a dual processor server with 2 quad core amd opterons and 16 gigabytes of ram.
7.) Yes 0 can occur, it is an artifact of the data acquisition system when it does but it can occur.
Your problem is the same as that faced by search engines. "I have a bajillion documents. I need the ones which contain this set of words." You just have (very conveniently), integers instead of words, and smallish documents. The solution is an inverted index. Introduction to Information Retrieval by Manning et al is (at that link) available free online, is very readable, and will go into a lot of detail about how to do this.
You're going to have to pay a price in disk space, but it can be parallelized, and should be more than fast enough to meet your timing requirements, once the index is constructed.
Assuming a random distribution of 0-16383, with a consistent 15 elements per set, and two billion sets, each element would appear in approximately 1.8M sets. Have you considered (and do you have the capacity for) building a 16384x~1.8M (30B entries, 4 bytes each) lookup table? Given such a table, you could query which sets contain (1) and (17) and (5555) and then find the intersections of those three ~1.8M-element lists.
My guess is as follows.
Assume that each set has a name or ID or address (a 4-byte number will do if there are only 2 billion of them).
Now walk through all the sets once, and create the following output files:
A file which contains the IDs of all the sets which contain '1'
A file which contains the IDs of all the sets which contain '2'
A file which contains the IDs of all the sets which contain '3'
... etc ...
If there are 16 entries per set, then on average each of these 2^16 files will contain the IDs of 2^20 sets; with each ID being 4 bytes, this would require 2^38 bytes (256 GB) of storage.
You'll do the above once, before you process requests.
When you receive requests, use these files as follows:
Look at a couple of numbers in the request
Open up a couple of the corresponding index files
Get the list of all sets which exist in both these files (there's only a million IDs in each file, so this should't be difficult)
See which of these few sets satisfy the remainder of the request
My guess is that if you do the above, creating the indexes will be (very) slow and handling requests will be (very) quick.
I have recently discovered methods that use Space Filling curves to map the multi-dimensional data down to a single dimension. One can then index the data based on its 1D index. Range queries can be easily carried out by finding the segments of the curve that intersect the box that represents the curve and then retrieving those segments.
I believe that this method is far superior to making the insane indexes as suggested because after looking at it, the index would be as large as the data I wished to store, hardly a good thing. A somewhat more detailed explanation of this can be found at:
http://www.ddj.com/184410998
and
http://www.dcs.bbk.ac.uk/~jkl/publications.html
Make 16383 index files, one for each possible search value. For each value in your input set, write the file position of the start of the set into the corresponding index file. It is important that each of the index files contains the same number for the same set. Now each index file will consist of ascending indexes into the master file.
To search, start reading the index files corresponding to each search value. If you read an index that's lower than the index you read from another file, discard it and read another one. When you get the same index from all of the files, that's a match - obtain the set from the master file, and read a new index from each of the index files. Once you reach the end of any of the index files, you're done.
If your values are evenly distributed, each index file will contain 1/16383 of the input sets. If your average search set consists of 6 values, you will be doing a linear pass over 6/16383 of your original input. It's still an O(n) solution, but your n is a bit smaller now.
P.S. Is zero an impossible result value, or do you really have 16384 possibilities?
Just playing devil's advocate for an approach which includes brute force + index lookup :
Create an index with the min , max and no of elements of sets.
Then apply brute force excluding sets where max < max(set being searched) and min > min (set being searched)
In brute force also exclude sets whole element count is less than that of the set being searched.
95% of your searches would really be brute forcing a very smaller subset. Just a thought.
We get these ~50GB data files consisting of 16 byte codes, and I want to find any code that occurs 1/2% of the time or more. Is there any way I can do that in a single pass over the data?
Edit: There are tons of codes - it's possible that every code is different.
EPILOGUE: I've selected Darius Bacon as best answer, because I think the best algorithm is a modification of the majority element he linked to. The majority algorithm should be modifiable to only use a tiny amount of memory - like 201 codes to get 1/2% I think. Basically you just walk the stream counting up to 201 distinct codes. As soon as you find 201 distinct codes, you drop one of each code (deduct 1 from the counters, forgetting anything that becomes 0). At the end, you have dropped at most N/201 times, so any code occurring more times than that must still be around.
But it's a two pass algorithm, not one. You need a second pass to tally the counts of the candidates. It's actually easy to see that any solution to this problem must use at least 2 passes (the first batch of elements you load could all be different and one of those codes could end up being exactly 1/2%)
Thanks for the help!
Metwally et al., Efficient Computation of Frequent and Top-k Elements in Data Streams (2005). There were some other relevant papers I read for my work at Yahoo that I can't find now; but this looks like a good start.
Edit: Ah, see this Brian Hayes article. It sketches an exact algorithm due to Demaine et al., with references. It does it in one pass with very little memory, yielding a set of items including the frequent ones you're looking for, if they exist. Getting the exact counts takes a (now-tractable) second pass.
this will depend on the distribution of the codes. if there are a small enough number of distinct codes you can build a http://en.wikipedia.org/wiki/Frequency_distribution in core with a map. otherwise you probably will have to build a http://en.wikipedia.org/wiki/Histogram and then make multiple passes over the data examining frequencies of codes in each bucket.
Sort chunks of the file in memory, as if you were performing and external sort. Rather than writing out all of the sorted codes in each chunk, however, you can just write each distinct code and the number of occurrences in that chunk. Finally, merge these summary records to find the number of occurrences of each code.
This process scales to any size data, and it only makes one pass over the input data. Multiple merge passes may be required, depending on how many summary files you want to open at once.
Sorting the file allows you to count the number of occurrences of each code using a fixed amount of memory, regardless of the input size.
You also know the total number of codes (either by dividing the input size by a fixed code size, or by counting the number of variable length codes during the sorting pass in a more general problem).
So, you know the proportion of the input associated with each code.
This is basically the pipeline sort * | uniq -c
If every code appears just once, that's no problem; you just need to be able to count them.
That depends on how many different codes exist, and how much memory you have available.
My first idea would be to build a hash table of counters, with the codes as keys. Loop through the entire file, increasing the counter of the respective code, and counting the overall number. Finally, filter all keys with counters that exceed (* overall-counter 1/200).
If the files consist solely of 16-byte codes, and you know how large each file is, you can calculate the number of codes in each file. Then you can find the 0.5% threshold and follow any of the other suggestions to count the occurrences of each code, recording each one whose frequency crosses the threshold.
Do the contents of each file represent a single data set, or is there an arbitrary cutoff between files? In the latter case, and assuming a fairly constant distribution of codes over time, you can make your life simpler by splitting each file into smaller, more manageable chunks. As a bonus, you'll get preliminary results faster and can pipeline then into the next process earlier.