Parquet storage size higher for duplicate data - apache-spark

I have a dataset which has close to 2 billion rows in parquet format which spans in 200 files. It occupies 17.4GB on S3. This dataset has close to 45% of duplicate rows. I deduplicated the dataset using 'distinct' function in Spark, and wrote it to a different location on S3.
I expected the data storage to be reduced by half. Instead, the deduplicated data took 34.4 GB (double of that which had duplicates).
I took to check the metadata of these two datasets. I found that there is a difference in the column encoding of the duplicate and deduplicated data.
Difference in column encodings
I want to understand how to get the expected behavior of reducing the storage size.
Having said that, I have a few questions further:
I also want to understand if this anomaly affect the performance in any way. In my process, I am having to do apply lot of filters on these columns and using distinct function while persisting the filtered data.
I have seen in a few parquet blogs online that encoding for a column is only one. In this case I see more than one column encodings. Is this normal?

Related

Spark dataframe distinct write is increasing the output size by almost 10 fold

I have a case where i am trying to write some results using dataframe write into S3 using the below query with input_table_1 size is 13 Gb and input_table_2 as 1 Mb
input_table_1 has columns account, membership and
input_table_2 has columns role, id , membership_id, quantity, start_date
SELECT
/*+ BROADCASTJOIN(input_table_2) */
account,
role,
id,
quantity,
cast(start_date AS string) AS start_date
FROM
input_table_1
INNER JOIN
input_table_2
ON array_contains(input_table_1.membership, input_table_2.membership_id)
where membership array contains list of member_ids
This dataset write using Spark dataframe is generating around 1.1TiB of data in S3 with around 700 billion records.
We identified that there are duplicates and used dataframe.distinct.write.parquet("s3path") to remove the duplicates . The record count is reduced to almost 1/3rd of the previous total count with around 200 billion rows but we observed that the output size in S3 is now 17.2 TiB .
I am very confused how this can happen.
I have used the following spark conf settings
spark.sql.shuffle.partitions=20000
I have tried to do a coalesce and write to s3 but it did not work.
Please suggest if this is expected and when can be done ?
There's two sides to this:
1) Physical translation of distinct in Spark
The Spark catalyst optimiser turns a distinct operation into an aggregation by means of the ReplaceDeduplicateWithAggregate rule (Note: in the execution plan distinct is named Deduplicate).
This basically means df.distinct() on all columns is translated into a groupBy on all columns with an empty aggregation:
df.groupBy(df.columns:_*).agg(Map.empty).
Spark uses a HashPartitioner when shuffling data for a groupBy on respective columns. Since the groupBy clause in your case contains all columns (well, implicitly, but it does), you're more or less randomly shuffling data to different nodes in the cluster.
Increasing spark.sql.shuffle.partitions in this case is not going to help.
Now on to the 2nd side, why does this affect the size of your parquet files so much?
2) Compression in parquet files
Parquet is a columnar format, will say your data is organised in columns rather than row by row. This allows for powerful compression if data is adequately laid-out & ordered. E.g. if a column contains the same value for a number of consecutive rows, it is enough to write that value just once and make a note of the number of repetitions (a strategy called run length encoding). But Parquet also uses various other compression strategies.
Unfortunately, data ends up pretty randomly in your case after shuffling to remove duplicates. The original partitioning of input_table_1 was much better fitted.
Solutions
There's no single answer how to solve this, but here's a few pointers I'd suggest doing next:
What's causing the duplicates? Could these be removed upstream? Or is there a problem with the join condition causing duplicates?
A simple solution is to just repartition the dataset after distinct to match the partitioning of your input data. Adding a secondary sorting (sortWithinPartition) is likely going to give you even better compression. However, this comes at the cost of an additional shuffle!
As #matt-andruff pointed out below, you can also achieve this in SQL using cluster by. Obviously, that also requires you to move the distinct keyword into your SQL statement.
Write your own deduplication algorithm as Spark Aggregator and group / shuffle the data just once in a meaningful way.

How do I most effectively compress my highly-unique columns?

I have a Spark DataFrame consisting of many double columns that are measurements, but I want a way of annotating each unique row by computing a hash of several other non-measurement columns. This hash results in garbled strings that are highly unique, and I've noticed my dataset size increases substantially when this column is present. How can I sort / lay out my data to decrease the overall dataset size?
I know that the Snappy compression protocol used on my parquet files executes best upon runs of similar data, so I think a sort over the primary key could be useful, but I also can't coalesce() the entire dataset into a single file (it's hundreds of GB in total size before the primary key creation step).
My hashing function is SHA2(128) FYI.
If you have a column that can be computed from the other columns, then simply omit that column before compression, and reconstruct it after decompression.

Specify max file size while write dataframe as parquet

When I try to write a dataframe as parquet, the file sizes are non-uniform. Although I don't want to make the files uniform, I want to set a max size for each file.
I can't afford to repartition the data as the dataframe is sorted(As per my understanding, repartitioning a sorted dataframe can distort the ordering).
Any help would be appreciated.
I have come across maxRecordsPerFile, but I don't want to limit the number of rows and I might not have full information about the columns(total number of columns and their types). So it's difficult to estimate file size based on rows.
I have read about parquet block size as well and I don't think that helps.

How does Parquet file size changes with the count in Spark Dataset

I came across a scenario where I had spark dataset with 24 columns, of which I was grouping by First 22 columns and summing up last two columns.
I removed the group by from the query and I have all 24 columns selected now.
Initial count of the dataset was 79,304.
After I removed group by the count has increased to 138,204 which is understood because I have removed the group by.
But I was not clear with the behavior that The initial size of parquet file was 2.3MB but later it got reduced to 1.5MB . Can anyone please help me understand this.
Also not every time the size reduces,
I had a similar scenario for 22 columns
count before was 35,298,226 and after removing group by was 59,874,208
and here the size has increased from 466.5MB to 509.8MB
When dealing with parquet sizes it's not about the number of rows it's about the data it self.
Parquet is columnar oriented format and therefore it store data column wise and compress the data column wise. Therefore it's not about the number of rows but rather the diversity of it's columns.
Parquet will do better compression as the diversity of the most diverse column in the table. So if you have one column dataframe it will be compress good as the distance between the values of the column.

pyspark: Efficiently have partitionBy write to same number of total partitions as original table

I had a question that is related to pyspark's repartitionBy() function which I originally posted in a comment on this question. I was asked to post it as a separate question, so here it is:
I understand that df.partitionBy(COL) will write all the rows with each value of COL to their own folder, and that each folder will (assuming the rows were previously distributed across all the partitions by some other key) have roughly the same number of files as were previously in the entire table. I find this behavior annoying. If I have a large table with 500 partitions, and I use partitionBy(COL) on some attribute columns, I now have for example 100 folders which each contain 500 (now very small) files.
What I would like is the partitionBy(COL) behavior, but with roughly the same file size and number of files as I had originally.
As demonstration, the previous question shares a toy example where you have a table with 10 partitions and do partitionBy(dayOfWeek) and now you have 70 files because there are 10 in each folder. I would want ~10 files, one for each day, and maybe 2 or 3 for days that have more data.
Can this be easily accomplished? Something like df.write().repartition(COL).partitionBy(COL) seems like it might work, but I worry that (in the case of a very large table which is about to be partitioned into many folders) having to first combine it to some small number of partitions before doing the partitionBy(COL) seems like a bad idea.
Any suggestions are greatly appreciated!
You've got several options. In my code below I'll assume you want to write in parquet, but of course you can change that.
(1) df.repartition(numPartitions, *cols).write.partitionBy(*cols).parquet(writePath)
This will first use hash-based partitioning to ensure that a limited number of values from COL make their way into each partition. Depending on the value you choose for numPartitions, some partitions may be empty while others may be crowded with values -- for anyone not sure why, read this. Then, when you call partitionBy on the DataFrameWriter, each unique value in each partition will be placed in its own individual file.
Warning: this approach can lead to lopsided partition sizes and lopsided task execution times. This happens when values in your column are associated with many rows (e.g., a city column -- the file for New York City might have lots of rows), whereas other values are less numerous (e.g., values for small towns).
(2) df.sort(sortCols).write.parquet(writePath)
This options works great when you want (1) the files you write to be of nearly equal sizes (2) exact control over the number of files written. This approach first globally sorts your data and then finds splits that break up the data into k evenly-sized partitions, where k is specified in the spark config spark.sql.shuffle.partitions. This means that all values with the same values of your sort key are adjacent to each other, but sometimes they'll span a split, and be in different files. This, if your use-case requires all rows with the same key to be in the same partition, then don't use this approach.
There are two extra bonuses: (1) by sorting your data its size on disk can often be reduced (e.g., sorting all events by user_id and then by time will lead to lots of repetition in column values, which aids compression) and (2) if you write to a file format the supports it (like Parquet) then subsequent readers can read data in optimally by using predicate push-down, because the parquet writer will write the MAX and MIN values of each column in the metadata, allowing the reader to skip rows if the query specifies values outside of the partition's (min, max) range.
Note that sorting in Spark is more expensive than just repartitioning and requires an extra stage. Behind the scenes Spark will first determine the splits in one stage, and then shuffle the data into those splits in another stage.
(3) df.rdd.partitionBy(customPartitioner).toDF().write.parquet(writePath)
If you're using spark on Scala, then you can write a customer partitioner, which can get over the annoying gotchas of the hash-based partitioner. Not an option in pySpark, unfortunately. If you really want to write a custom partitioner in pySpark, I've found this is possible, albeit a bit awkward, by using rdd.repartitionAndSortWithinPartitions:
df.rdd \
.keyBy(sort_key_function) \ # Convert to key-value pairs
.repartitionAndSortWithinPartitions(numPartitions=N_WRITE_PARTITIONS,
partitionFunc=part_func) \
.values() # get rid of keys \
.toDF().write.parquet(writePath)
Maybe someone else knows an easier way to use a custom partitioner on a dataframe in pyspark?
df.repartition(COL).write().partitionBy(COL)
will write out one file per partition. This will not work well if one of your partition contains a lot of data. e.g. if one partition contains 100GB of data, Spark will try to write out a 100GB file and your job will probably blow up.
df.repartition(2, COL).write().partitionBy(COL)
will write out a maximum of two files per partition, as described in this answer. This approach works well for datasets that are not very skewed (because the optimal number of files per partition is roughly the same for all partitions).
This answer explains how to write out more files for the partitions that have a lot of data and fewer files for the small partitions.

Resources