Really stuck on this! Assume I have a following data set:
A | B
------------------
1/2/12 | 13:3.8
04:4.1 | 12:1.4
15:4.3 | 1/3/13
Observations A and B are in general in the format minutes:seconds.milliseconds like A is a click and B is a response. Sometimes time format has a form of month/day/year if any of the events happens to be in the beginning of the new day.
What I want? Is to calculate average difference between B and A. I can easily handle m:s.ms as splitting them into two parts for each A and B and then cast as DOUBLE and perform all needed operations but it all fails when m/d/yy are introduced. The easiest way to omit them but it is not a really good practice. Is there is a clear way to handle such exceptions using PIG?
A thought worth contemplating ....
Ref : http://pig.apache.org/docs/r0.12.0/func.html for String and Date functions used.
Input :
1/2/12|13:3.8
04:4.1|12:1.4
15:4.3|1/3/13
Pig Script :
A = LOAD 'input.csv' USING PigStorage('|') AS (start_time:chararray,end_time:chararray);
B = FOREACH A GENERATE (INDEXOF(end_time,'/',0) > 0 AND LAST_INDEX_OF(end_time,'/') > 0 AND (INDEXOF(end_time,'/',0) != LAST_INDEX_OF(end_time,'/'))
? (ToUnixTime(ToDate(end_time,'MM/dd/yy'))) : (ToUnixTime(ToDate(end_time,'mm:ss.S')))) -
(INDEXOF(start_time,'/',0) >0 AND LAST_INDEX_OF(start_time,'/') > 0 AND (INDEXOF(start_time,'/',0) != LAST_INDEX_OF(start_time,'/'))
? (ToUnixTime(ToDate(start_time,'MM/dd/yy'))) : (ToUnixTime(ToDate(start_time,'mm:ss.S')))) AS diff_time;
C = FOREACH (GROUP B ALL) GENERATE AVG(B.diff_time);
DUMP C;
N.B. In place of ToUnixTime we can use ToMilliSeconds() method.
Output :
(1.0569718666666666E7)
Related
I am trying to build a pig script that takes in a textbook file and divides it into chapters and then compares the words in each chapter and returns only words that show up in all chapters and counts them. The chapters are Delimited fairly easily by CHAPTER - X.
Here's what I have so far:
lines = LOAD '../../Alice.txt' AS (line:chararray);
lineswithoutspecchars = FOREACH lines GENERATE REPLACE(line,'([^a-zA-Z\\s]+)','') as line;
words = FOREACH lineswithoutspecchars GENERATE FLATTEN(TOKENIZE(line)) as word;
grouped = GROUP words BY word;
wordcount = FOREACH grouped GENERATE group, COUNT(words);
DUMP wordcount;
Sorry that this question is probably way too simple compared to what I normally ask on stackoverflow and I googled around for it but perhaps I am not using the correct keywords. I am brand new to PIG and trying to learn it for a new job assignment.
Thanks in advance!
A bit lengthy but you will get the result. You could cut down unnecessary relations based on your file though. Provided appropriate comments in teh script.
Input File:
Pig does not know whether integer values in baseball are stored as ASCII strings, Java
serialized values, binary-coded decimal, or some other format. So it asks the load func-
tion, because it is that function’s responsibility to cast bytearrays to other types. In
general this works nicely, but it does lead to a few corner cases where Pig does not know
how to cast a bytearray. In particular, if a UDF returns a bytearray, Pig will not know
how to perform casts on it because that bytearray is not generated by a load function.
CHAPTER - X
In a strongly typed computer language (e.g., Java), the user must declare up front the
type for all variables. In weakly typed languages (e.g., Perl), variables can take on values
of different type and adapt as the occasion demands.
CHAPTER - X
In this example, remember we are pretending that the values for base_on_balls and
ibbs turn out to be represented as integers internally (that is, the load function con-
structed them as integers). If Pig were weakly typed, the output of unintended would
be records with one field typed as an integer. As it is, Pig will output records with one
field typed as a double. Pig will make a guess and then do its best to massage the data
into the types it guessed.
Pig Script:
A = LOAD 'file' as (line:chararray);
B = FOREACH A GENERATE REPLACE(line,'([^a-zA-Z\\s]+)','') as line;
//we need to split on CHAPTER X but the above load function would give us a tuple for each newline. so
group everything and convert that bag to string which will give a single tuple with _ as delimiter.
C = GROUP B ALL;
D = FOREACH C GENERATE BagToString(B) as (line:chararray);
//now we dont have any commas so convert our delimiter CHAPTER X to comma. We do this becuase if we pass this
to TOKENIZE it would split that into separarte column that would be useful to RANK it.
E = FOREACH D GENERATE REPLACE(line,'_CHAPTER X_',',') AS (line:chararray);
F = FOREACH E GENERATE REPLACE(line,'_',' ') AS (line:chararray); //remove the delimiter created by BagToString
//create separate columns
G = FOREACH F GENERATE FLATTEN(TOKENIZE(line,',')) AS (line:chararray);
//we need to rank each chapter so that would be easy when you are doing the count of each word.
H = RANK G;
J = FOREACH H GENERATE rank_G,FLATTEN(TOKENIZE(line)) as (line:chararray);
J1 = GROUP J BY (rank_G, line);
J2 = FOREACH J1 GENERATE COUNT(J) AS (cnt:long),FLATTEN(group.line) as (word:chararray),FLATTEN(group.rank_G) as (rnk:long);
//So J2 result will not have duplicate word within each chapter now.
//So if we group it by word and then filter teh count of that by 2 we are sure that the word is present in all chapters.
J3 = GROUP J2 BY word;
J4 = FOREACH J3 GENERATE SUM(J2.cnt) AS (sumval:long),COUNT(J2) as (cnt:long),FLATTEN(group) as (word:chararray);
J5 = FILTER J4 BY cnt > 2;
J6 = FOREACH J5 GENERATE word,sumval;
dump J6;
//result in order word,count across chapters
Output:
(a,8)
(In,5)
(as,6)
(the,9)
(values,4)
Is there any possibility to cut a certain area after the decimal point of a float or double number?
For example: the result would be 2.67894 => I want to have 2.6 as result (and not 2.7 when rounded).
try it.. val is your values like 2.666,3.666,4.666666,5.3456334.....
b = foreach a GENERATE (FLOOR(val * 10) / 10);
dump b;
Write a UDF (User Defined Function) for this.
A very simple python UDF (numformat.py):
#outputSchema('value:double')
def format(data):
return round(data,1)
(Of course you can parametrized the UDF to use different precision.)
Than register and use it in your pig code. Example:
REGISTER numformat.py USING jython as numformat;
A = LOAD 'so/testdata.csv' USING PigStorage(',') AS (data:double);
B = FOREACH A GENERATE numformat.format(data);
DUMP B;
For the following input:
2.1234
12.334
The dumped result is:
(2.1)
(12.3)
Split the group record in to different records :
for eg :
Input : (A,(3,2,3))
Output in to 3 new lines:
A,3
A,2
A,3
Can any one let me know the option to do this please?
The problem is when you convert the output of Arraylist to tuple then it will be difficult to achieve what you want, so I recommend this approach, so it will be easy to get the output .
In your UDF code, instead of creating Arraylist, append the output into string seperated by comma and return back to pig script.
You final output should be like this from UDF as a string ie "3,2,3"
Then use the below code to get the result
C = FOREACH B GENERATE $0,NewRollingCount(BagToString($1)) AS rollingCnt
D = FOREACH C GENERATE $0,FLATTEN(TOKENIZE(rollingcnt));
DUMP D;
I'm a beginner in Talend Open Studio, and I'm trying to do the transformation below.
From a SQL Table that contains:
DeltaStock Date
------------------------
+50 (initial stock) J0
+80 J1
-30 J2
... ...
I want to produce this table:
Stock Date
-----------
50 J0
130 J1
100 J2
... ...
Do you think this could be possible using TOS? I thought of using tAggregateRow, but I didn't find it appropriate to my issue.
There's probably an easier way to do this using the tMemorizeRows component but the first thought that comes to mind is to use the globalMap to store a rolling sum.
In Talend it is possible to store an object (any value or any type) in the globalMap so that it can be retrieved later on in the job. This is used automatically if you ever use a tFlowToIterate component which allows you to retrieve the values for that row that is being iterated on from the globalMap.
A very basic sample job might look like this:
In this we have a tJava component that only initialises the rolling sum in the globalMap with the following code:
//Initialise the rollingSum global variable
globalMap.put("rollingSum", 0);
After this we connect this component onSubjobOk to make sure we only carry on if we've managed to put the rollingSum into the globalMap.
I then provide my data using a tFixedFlowInput component which allows me to easily hardcode some values for this example job. You could easily replace this with any input. I have used your sample input data from the question:
We then process the data using a tJavaRow which will do some transformations on the data row by row. I've used the following code which works for this example:
//Initialise the operator and the value variables
String operator = "";
Integer value = 0;
//Get the current rolling sum
Integer rollingSum = (Integer) globalMap.get("rollingSum");
//Extract the operator
Pattern p = Pattern.compile("^([+-])([0-9]+)$");
Matcher m = p.matcher(input_row.deltaStock);
//If we have any matches from the regular expression search then extract the operator and the value
if (m.find()) {
operator = m.group(1);
value = Integer.parseInt(m.group(2));
}
//Conditional to use the operator
if ("+".equals(operator)) {
rollingSum += value;
} else if ("-".equals(operator)) {
rollingSum -= value;
} else {
System.out.println("The operator provided wasn't a + or a -");
}
//Put the new rollingSum back into the globalMap
globalMap.put("rollingSum", rollingSum);
//Output the data
output_row.stock = rollingSum;
output_row.date = input_row.date;
There's quite a lot going on there but basically it starts by getting the current rollingSum from the globalMap.
Next, it uses a regular expression to split up the deltaStock string into an operator and a value. From this it uses the operator provided (plus or minus) to either add the deltaStock to the rollingSum or subtract the deltaStock from the rollingSum.
After this it then adds the new rollingSum back into the globalMap and outputs the 2 columns of stock and date (unchanged).
In my sample job I then output the data using a tLogRow which will print the values of the data to the console. I typically select the table formatting option in it and in this case I get the following output:
.-----+----.
|tLogRow_8 |
|=----+---=|
|stock|date|
|=----+---=|
|50 |J0 |
|130 |J1 |
|100 |J2 |
'-----+----'
Which should be what you were looking for.
You should be able to do it in Talend Open Studio.
I attach here an image with the JOB, the content of the tJavaRow and the execution result.
I left under the tFixedFlowInput used to simulate the input a tJDBCInput that you should use to read the data from your DB. Hopefully you can use a specific tXXXInput for your DB instead of the generic JDBC one.
Here is some simple code in the tJavaRow.
//Code generated according to input schema and output schema
output_row.delta = input_row.delta;
output_row.date = input_row.date;
output_row.rollingSum =
Integer.parseInt(globalMap.get("rollingSum").toString());
int delta = Integer.parseInt(input_row.delta);
output_row.rollingSum += delta;
// Save rolling SUM for next round
globalMap.put("rollingSum", output_row.rollingSum);
Beware of the exceptions in the parseInt(). You should handle them the way you feel right.
In my projects I usually have a SafeParse library that does not throws exceptions but returns a default value I can pass together with the vale to be parsed.
I have input records of the form
2013-07-09T19:17Z,f1,f2
2013-07-09T03:17Z,f1,f2
2013-07-09T21:17Z,f1,f2
2013-07-09T16:17Z,f1,f2
2013-07-09T16:14Z,f1,f2
2013-07-09T16:16Z,f1,f2
2013-07-09T01:17Z,f1,f2
2013-07-09T16:18Z,f1,f2
These represent timestamps and events. I have written these by hand, but actual data should be sorted based on time.
I would like to generate a set of records which would be input to graph plotting function which needs continuous time series. I would like to fill in missing values, i.e. if there are entries for "2013-07-09T19:17Z" and "2013-07-09T19:19Z", I would like to generate entry for "2013-07-09T19:18Z" with predefined value.
My thoughts on doing this:
Use MIN and MAX to find the start and end date in the series
Write UDF which takes min and max and returns relation with missing
timestamps
Join above 2 relations
I cannot get my head around on how to implement this in PIG though. Would appreciate any help.
Thanks!
Generate another file using a script (outside pig)with all time stamps between MIN and MAX , including MIN and MAX. Load this as a second data set. Here is a sample that I used from your data set. Please note I filled in only few gaps not all.
2013-07-09T01:17Z,d1,d2
2013-07-09T01:18Z,d1,d2
2013-07-09T03:17Z,d1,d2
2013-07-09T16:14Z,d1,d2
2013-07-09T16:15Z,d1,d2
2013-07-09T16:16Z,d1,d2
2013-07-09T16:17Z,d1,d2
2013-07-09T16:18Z,d1,d2
2013-07-09T19:17Z,d1,d2
2013-07-09T21:17Z,d1,d2
Do a COGROUP on the original dataset and the generated dataset above. Use a nested FOREACH GENERATE to write output dataset. If first dataset is empty, use the values from second set to generate output dataset else the first dataset. Here is the piece of code I used on these two datasets.
Org_Set = LOAD 'pigMissingData/timeSeries' USING PigStorage(',') AS (timeStamp, fl1, fl2);
Default_set = LOAD 'pigMissingData/timeSeriesFull' USING PigStorage(',') AS (timeStamp, fl1, fl2);
coGrouped = COGROUP Org_Set BY timeStamp, Default_set BY timeStamp;
Filled_Data_set = FOREACH coGrouped {
x = COUNT(times);
y = (x == 0? (Default_set.fl1, Default_set.fl2): (Org_Set.fl1, Org_Set.fl2));
GENERATE FLATTEN(group), FLATTEN(y.$0), FLATTEN(y.$1);
};
if you need further clarification or help let me know
In addition to #Rags answer, you could use the STREAM x THROUGH command and a simple awk script (similar to this one) to generate the date range once you have the min and max dates. Something similar to (untested! - you might need to single line the awk script with semi-colon command delimitation, or better to ship it as a script file)
grunt> describe bounds;
(min:chararray, max:chararray)
grunt> dump bounds;
(2013/01/01,2013/01/04)
grunt> fullDateBounds = STREAM bounds THROUGH `gawk '{
split($1,s,"/")
split($2,e,"/")
st=mktime(s[1] " " s[2] " " s[3] " 0 0 0")
et=mktime(e[1] " " e[2] " " e[3] " 0 0 0")
for (i=st;i<=et;i+=60*24) print strftime("%Y/%m/%d",i)
}'`;