I'm very new to Pig so I may be going about this the wrong way. I have a bunch of gzipped files in a directory in Hadoop. I'm trying to sample around 1000 lines from all of these files put together. It doesn't have to be exact, so I wanted to use SAMPLE. SAMPLE needs a probability of sampling a line, rather than the number of lines that I need, so I thought I should count up the number of lines among all these files and than simply divide 1000 by that count and use it as the probability. This will work, since I don't need to have exactly 100 lines at the end. Here is what I got so far:
raw = LOAD '/data_dir';
cnt = FOREACH (GROUP raw ALL) GENERATE COUNT_STAR(raw);
cntdiv = FOREACH cnt GENERATE (float)100/ct.$0;
Now I'm not sure how to use the value in cntdiv in SAMPLE. I tried SAMPLE raw cntdiv and SAMPLE raw cntdiv.$0, but they don't work. Can I even use that value in the call to SAMPLE? Maybe there is a much better way of accomplishing what I'm trying to do?
Check out the description in the ticket originally requesting this feature: https://issues.apache.org/jira/browse/PIG-1926
I haven't tested this, but it looks like this should work:
raw = LOAD '/data_dir';
samplerate = FOREACH (GROUP raw ALL) GENERATE 1000.0/COUNT_STAR(raw) AS rate;
thousand = SAMPLE raw samplerate.rate;
The important thing is to refer to your scalar by name (rate), not by position ($0).
Related
I am trying to reduce the size of my data and I cannot make it work. I have data points taken every minute over 1 month. I want to reduce this data to have one sample for every hour. The problem is: Some of my runs have "NA" value, so I delete these rows. There is not exactly 60 points for every hour - it varies.
I have a 'Timestamp' column. I have used this to make a 'datehour' column which has the same value if the data set has the same date and hour. I want to average all the values with the same 'datehour' value.
How can I do this? I have tried using the if and for loop below, but it takes so long to run.
Thanks for all your help! I am new to Julia and come from a Matlab background.
======= CODE ==========
uniquedatehour=unique(datehour,1)
index=[]
avedata=reshape([],0,length(alldata[1,:]))
for j in uniquedatehour
for i in 1:length(datehour)
if datehour[i]==j
index=vcat(index,i)
else
rows=alldata[index,:]
rows=convert(Array{Float64,2},rows)
avehour=mean(rows,1)
avedata=vcat(avedata,avehour)
index=[]
continue
end
end
end
There are several layers to optimizing this code. I am assuming that your data is sorted on datehour (your code assumes this).
Layer one: general recommendation
Wrap your code in a function. Executing code in global scope in Julia is much slower than within a function. By wrapping it make sure to either pass data to your function as arguments or if data is in global scope it should be qualified with const;
Layer two: recommendations to your algorithm
Statement like [] creates an array of type Any which is slow, you should use type qualifier like index=Int[] to make it fast;
Using vcat like index=vcat(index,i) is inefficient, it is better to do push!(index, i) in place;
It is better to preallocate avedata with e.g. fill(NA, length(uniquedatehour), size(alldata, 2)) and assign values to an existing matrix than to do vcat on it;
Your code will produce incorrect results if I am not mistaken as it will not catch the last entry of uniquedatehour vector (assume it has only one element and check what happens - avedata will have zero rows)
Line rows=convert(Array{Float64,2},rows) is probably not needed at all. If alldata is not Matrix{Float64} it is better to convert it at the beginning with Matrix{Float64}(alldata);
You can change line rows=alldata[index,:] to a view like view(alldata, index, :) to avoid allocation;
In general you can avoid creation of index vector as it is enough that you remember start s and end e position of the range of the same values and then use range s:e to select rows you want.
If you correct those things please post your updated code and maybe I can help further as there is still room for improvement but requires a bit different algorithmic approach (but maybe you will prefer option below for simplicity).
Layer three: how I would do it
I would use DataFrames package to handle this problem like this:
using DataFrames
df = DataFrame(alldata) # assuming alldata is Matrix{Float64}, otherwise convert it here
df[:grouping] = datehour
agg = aggregate(df, :grouping, mean) # maybe this is all what you need if DataFrame is OK for you
Matrix(agg[2:end]) # here is how you can convert DataFrame back to a matrix
This is not the fastest solution (as it converts to a DataFrame and back but it is much simpler for me).
I have written a Python 2.7 script that reads a CSV file and then does some standard deviation calculations . It works absolutely fine however it is very very slow. A CSV I tried with 100 million lines took around 28 hours to complete. I did some googling and it appears that maybe using the pandas module might makes this quicker .
I have posted part of the code below, since i am a pretty novice when it comes to python , i am unsure if using pandas would actually help at all and if it did would the function need to be completely re-written.
Just some context for the CSV file, it has 3 columns, first column is an IP address, second is a url and the third is a timestamp.
def parseCsvToDict(filepath):
with open(csv_file_path) as f:
ip_dict = dict()
csv_data = csv.reader(f)
f.next() # skip header line
for row in csv_data:
if len(row) == 3: #Some lines in the csv have more/less than the 3 fields they should have so this is a cheat to get the script working ignoring an wrong data
current_ip, URI, current_timestamp = row
epoch_time = convert_time(current_timestamp) # convert each time to epoch
if current_ip not in ip_dict.keys():
ip_dict[current_ip] = dict()
if URI not in ip_dict[current_ip].keys():
ip_dict[current_ip][URI] = list()
ip_dict[current_ip][URI].append(epoch_time)
return(ip_dict)
Once the above function has finished the data is parsed to another function that calculates the standard deviation for each IP/URL pair (using numpy.std).
Do you think that using pandas may increase the speed and would it require a complete rewrite or is it easy to modify the above code?
The following should work:
import pandas as pd
colnames = ["current_IP", "URI", "current_timestamp", "dummy"]
df = pd.read_csv(filepath, names=colnames)
# Remove incomplete and redundant rows:
df = df[~df.current_timestamp.isnull() & df.dummy.isnull()]
Notice this assumes you have enough RAM. In your code, you are already assuming you have enough memory for the dictionary, but the latter may be significatively smaller than the memory used by the above, for two reasons.
If it is because most lines are dropped, then just parse the csv by chunks: arguments skiprows and nrows are your friends, and then pd.concat
If it is because IPs/URLs are repeated, then you will want to transform IPs and URLs from normal columns to indices: parse by chunks as above, and on each chunk do
indexed = df.set_index(["current_IP", "URI"]).sort_index()
I expect this will indeed give you a performance boost.
EDIT: ... including a performance boost to the calculation of the standard deviation (hint: df.groupby())
I will not be able to give you an exact solution, but here are a couple of ideas.
Based on your data, you read 100000000. / 28 / 60 / 60 approximately 1000 lines per second. Not really slow, but I believe that just reading such a big file can cause a problem.
So take a look at this performance comparison of how to read a huge file. Basically a guy suggests that doing this:
file = open("sample.txt")
while 1:
lines = file.readlines(100000)
if not lines:
break
for line in lines:
pass # do something
can give you like 3x read boost. I also suggest you to try defaultdict instead of your if k in dict create [] otherwise append.
And last, not related to python: working in data-analysis, I have found an amazing tool for working with csv/json. It is csvkit, which allows to manipulate csv data with ease.
In addition to what Salvador Dali said in his answer: If you want to keep as much of the current code of your script, you may find that PyPy can speed up your program:
“If you want your code to run faster, you should probably just use PyPy.” — Guido van Rossum (creator of Python)
I need logic for below scenario which needs to be implemented using Pig scripts. Can anyone please help in providing some ideas on how to do this.
Input contains a column groupName with some data like others and unknown. This data needs to be replaced by its previous record data.
Input:
id,groupName
123,casc0001
124,casc0002
125,sale0001
126,unknown
127,nave9876
128,casc0001
129,sale0002
130,others
131,casc0004
132,unknown
133,unknown
134,others
135,nave1234
output:
123,casc0001
124,casc0002
125,sale0001
126,sale0001
127,nave9876
128,casc0001
129,sale0002
130,sale0002
131,casc0004
132,casc0004
133,casc0004
134,casc0004
135,nave1234
In the above input 126,unknown to be replaced with 125,sale0001. 130,others need to be replaced by 129,sale0002. 132,unknown 133,unknown 134,others to be replaced with 131,casc0004.
--Edit--
I tried lead function in Pig. But it is used only to compare n rows at a time. Which cannot solve this completely.
Another logic which is working, but looking for optimized one.
Cogroup for the same data set (like Dataset and Dataset_self)
-Filter Dataset.id=Dataset_self.id or Dataset_self.groupname='others' or Dataset_self.groupname='unknown'
-Generate IdDiff like (Dataset_self.id-Dataset.id), CASE when id=id then ( id, group) else (id_self,group)
-Foreach (group id){
ordered = order by id,diff,group;
limited = ordered limit 1;
generate limited ;
}
This is going to be a complicated problem on a distributed system like hadoop, especially that your file is going to be split between nodes. In your case what if 126 happens to be the first record in a new split. Then you will need to trace the previous file split which is most likely on a different node. Lets say you come up with a MapReduce program to do this, in all likelyhood it would an extremely slow and inefficient way to do it. The solution might be simpler if you are in a single node system where the splittable property of your input format is false, and the nuber of reducers is set to 1.
In that case you could almost make the argument that a traditional database like Oracle or Terra data might be a better fit for your problem as you have lead or lag functions readily available which could be used to do exactly what u need.
In pig I massaged my data into something like:
(a,{(b,c),(d,e),(f,g)})
(h,{(i,j),(k,l)})
where the first item is the group and the bag are other values related to the group. I would like to get it into the following format:
(a,b,c,d,e,f,g)
(h,i,j,k,l)
I got to where I am now with
grunt> j = foreach G {
>> o = order myvar by second;
>> generate group, o.(first,second);
>> };
So the tuples in the bag are currently ordered. If I do something like mystuff = foreach j generate group, flatten($1); I get it all flattened and un-grouped.
Is this possible in pig, and if so what command should I be looking at?
There is no way I can that can do what you want out of the box. You really need to use a user-defined function for this. I know it sucks because you have to write Java or Python code, but you'll find several situations where Pig just doesn't go far enough. Pig can be considered a data flow language and not so much of a programming language, which is why UDFs play such an important role: they bridge the gap.
My suggestion is you write a UDF that takes in the group and value bag as parameters. Do the ordering/sorting in the UDF and also the flattening.
The other thing you want to be careful about is that now your rows will have different numbers of columns and Pig doesn't really like this. If you are just immediately outputting it, you can probably get away with this. You might want to consider having your UDF write out the list in a tab-delimited string or something that is preformatted. This isn't that big of a deal... feel free to ignore my advice here.
I do have to deal with very large plain text files (over 10 gigabytes, yeah I know it depends what we should call large), with very long lines.
My most recent task involves some line editing based on data from another file.
The data file (which should be modified) contains 1500000 lines, each of them are e.g. 800 chars long. Each line is unique, and contains only one identity number, each identity number is unique)
The modifier file is e.g. 1800 lines long, contains an identity number, and an amount and a date which should be modified in the data file.
I just transformed (with Vim regex) the modifier file to sed, but it's very inefficient.
Let's say I have a line like this in the data file:
(some 500 character)id_number(some 300 character)
And I need to modify data in the 300 char part.
Based on the modifier file, I come up with sed lines like this:
/id_number/ s/^\(.\{650\}\).\{20\}/\1CHANGED_AMOUNT_AND_DATA/
So I have 1800 lines like this.
But I know, that even on a very fast server, if I do a
sed -i.bak -f modifier.sed data.file
It's very slow, because it has to read every pattern x every line.
Isn't there a better way?
Note: I'm not a programmer, had never learnt (in school) about algorithms.
I can use awk, sed, an outdated version of perl on the server.
My suggested approaches (in order of desirably) would be to process this data as:
A database (even a simple SQLite-based DB with an index will perform much better than sed/awk on a 10GB file)
A flat file containing fixed record lengths
A flat file containing variable record lengths
Using a database takes care of all those little details that slow down text-file processing (finding the record you care about, modifying the data, storing it back to the DB). Take a look for DBD::SQLite in the case of Perl.
If you want to stick with flat files, you'll want to maintain an index manually alongside the big file so you can more easily look up the record numbers you'll need to manipulate. Or, better yet, perhaps your ID numbers are your record numbers?
If you have variable record lengths, I'd suggest converting to fixed-record lengths (since it appears only your ID is variable length). If you can't do that, perhaps any existing data will not ever move around in the file? Then you can maintain that previously mentioned index and add new entries as necessary, with the difference is that instead of the index pointing to record number, you now point to the absolute position in the file.
I suggest you a programm written in Perl (as I am not a sed/awk guru and I don't what they are exactly capable of).
You "algorithm" is simple: you need to construct, first of all, an hashmap which could give you the new data string to apply for each ID. This is achieved reading the modifier file of course.
Once this hasmap in populated you may browse each line of your data file, read the ID in the middle of the line, and generate the new line as you've described above.
I am not a Perl guru too , but I think that the programm is quite simple. If you need help to write it, ask for it :-)
With perl you should use substr to get id_number, especially if id_number has constant width.
my $id_number=substr($str, 500, id_number_length);
After that if $id_number is in range, you should use substr to replace remaining text.
substr($str, -300,300, $new_text);
Perl's regular expressions are very fast, but not in this case.
My suggestion is, don't use database. Well written perl script will outperform database in order of magnitude in this sort of task. Trust me, I have many practical experience with it. You will not have imported data into database when perl will be finished.
When you write 1500000 lines with 800 chars it seems 1.2GB for me. If you will have very slow disk (30MB/s) you will read it in a 40 seconds. With better 50 -> 24s, 100 -> 12s and so. But perl hash lookup (like db join) speed on 2GHz CPU is above 5Mlookups/s. It means that your CPU bound work will be in seconds and you IO bound work will be in tens of seconds. If it is really 10GB numbers will change but proportion is same.
You have not specified if data modification changes size or not (if modification can be done in place) thus we will not assume it and will work as filter. You have not specified what format of your "modifier file" and what sort of modification. Assume that it is separated by tab something like:
<id><tab><position_after_id><tab><amount><tab><data>
We will read data from stdin and write to stdout and script can be something like this:
my $modifier_filename = 'modifier_file.txt';
open my $mf, '<', $modifier_filename or die "Can't open '$modifier_filename': $!";
my %modifications;
while (<$mf>) {
chomp;
my ($id, $position, $amount, $data) = split /\t/;
$modifications{$id} = [$position, $amount, $data];
}
close $mf;
# make matching regexp (use quotemeta to prevent regexp meaningful characters)
my $id_regexp = join '|', map quotemeta, keys %modifications;
$id_regexp = qr/($id_regexp)/; # compile regexp
while (<>) {
next unless m/$id_regexp/;
next unless $modifications{$1};
my ($position, $amount, $data) = #{$modifications{$1}};
substr $_, $+[1] + $position, $amount, $data;
}
continue { print }
On mine laptop it takes about half minute for 1.5 million rows, 1800 lookup ids, 1.2GB data. For 10GB it should not be over 5 minutes. Is it reasonable quick for you?
If you start think you are not IO bound (for example if use some NAS) but CPU bound you can sacrifice some readability and change to this:
my $mod;
while (<>) {
next unless m/$id_regexp/;
$mod = $modifications{$1};
next unless $mod;
substr $_, $+[1] + $mod->[0], $mod->[1], $mod->[2];
}
continue { print }
You should almost certainly use a database, as MikeyB suggested.
If you don't want to use a database for some reason, then if the list of modifications will fit in memory (as it currently will at 1800 lines), the most efficient method is a hashtable populated with the modifications as suggested by yves Baumes.
If you get to the point where even the list of modifications becomes huge, you need to sort both files by their IDs and then perform a list merge -- basically:
Compare the ID at the "top" of the input file with the ID at the "top" of the modifications file
Adjust the record accordingly if they match
Write it out
Discard the "top" line from whichever file had the (alphabetically or numerically) lowest ID and read another line from that file
Goto 1.
Behind the scenes, a database will almost certainly use a list merge if you perform this alteration using a single SQL UPDATE command.
Good deal on the sqlloader or datadump decision. That's the way to go.