I need to write to Hadoop around 150B nested Json records per day (Using Spark Java),
What is the "fast" way to do it in term of performance for example :
Create Hive table and write parquet file to HDFS
Or create Dataset from Json and using saveAsTable
Or is there another way of doing it ?
Related
I have a question, is it possible to execute ETL for data using flume.
To be more specific I have flume configured on spoolDir which contains CSV files and I want to convert those files into Parquet files before storing them into Hadoop. Is it possible ?
If it's not possible would you recommend transforming them before storing in Hadoop or transform them using spark on Hadoop?
I'd probably suggest using nifi to move the files around. Here's a specific tutorial on how to do that with Parquet. I feel nifi was the replacement for Apache Flume.
Flume partial answers:(Not Parquet)
If you are flexible on format you can use an avro sink. You can use a hive sink and it will create a table in ORC format.(You can see if it also allows parquet in the definition but I have heard that ORC is the only supported format.)
You could likely use some simple script to use hive to move the data from the Orc table to a Parquet table. (Converting the files into the parquet files you asked for.)
Assuming I have an external hive table on top parquet/orc files partitioned on date, what would be the performance impact of using
spark.read.parquet("s3a://....").filter("date_col='2021-06-20'")
v/s
spark.sql("select * from table").filter("date_col='2021-06-20'")
After reading into a dataframe, It will be followed by a series of transformations and aggregations.
spark version : 2.3.0 or 3.0.2
hive version : 1.2.1000
number of records per day : 300-700 Mn
My hunch is that there won't be any performance difference while using either of the above queries since parquet natively has most of the optimizations that a hive metastore can provide and spark is capable of using it. Like, predicate push-down, advantages of columnar storage etc.
As a follow-up question, what happens if
The underlying data was csv instead of parquet. Does having a hive table on top improves performance ?
Hive table was bucketed. Does it make sense to read the underlying file system in this case instead of reading from table ?
Also, are there any situations where reading directly from parquet is a better option compared to hive ?
Hive should actually be faster here because they both have pushdowns, Hive already has the schema stored. The parquet read as you have it here will need to infer the merged schema. You can make them about the same by providing the schema.
You can make the Parquet version even faster by navigating directly to the partition. This avoids having to do the initial filter on the available partitions.
So something like this would do it:
spark.read.option("basePath", "s3a://....").parquet("s3a://..../date_col=2021-06-20")
Note this works best if you already have a schema, because this also skips schema merging.
As to your follow-ups:
It would make a huge difference if it's CSV, as it would then have to parse all of the data and then filter out those columns. CSV is really bad for large datasets.
Shouldn't really gain you all that much and may get you into trouble. The metadata that Hive stores can allow Spark to navigate your data more efficiently here than you trying to do it yourself.
What would be the most memory efficient way to copy the contents of a large relational table using spark and then write to a partitioned Hive table in parquet format (without sqoop). I have a basic spark app and i have done some other tuning with spark's jdbc but data in relational table is still 0.5 TB and 2 Billion records so I although I can lazy load the full table, I'm trying to figure out how to efficiently partition by date and save to hdfs without running into memory issues. since the jdbc load() from spark will load everything into memory I was thinking of looping through the dates in the database query but still not sure how to make sure I don't run out of memory.
If you need to use Spark you can add to your application date parameter for filtering table by date and run your Spark application in loop for each date. You can use bash or other scripting language for this loop.
This can look like:
foreach date in dates
spark-submit your application with date parameter
read DB table with spark.read.jdbc
filter by date using filter method
write result to HDFS with df.write.parquet("hdfs://path")
Another option is to use different technology for example implement Scala application using JDBC and DB cursor to iterate through rows and save result to HDFS. This is more complex, because you need to solve problems related to writing to Parquet format and saving to HDFS using Scala. If you want I can provide Scala code responsible for writing to Parquet format.
I have a hive table (80 million records) with the followig schema (event_id ,country,unit_id,date) and i need to export this data to a text file as with the following requirments:
1-Rows are aggregated(combined) by event_id.
2-Aggregated rows must be sorted according to date.
For example rows with same event_id must be combined as a list of lists, ordered according to date.
What is the best performance wise solution to make this job using spark ?
Note: This is expected to be a batch job.
Performance-wise, I think the best solution is to write a spark program (scala or python) to read in the underlying files to the hive table, do your transformations, and then write the output as a file.
I've found that it's much quicker to just read the files in spark rather than querying hive through spark and pulling the result into a dataframe.
I am trying to generate parquet files in S3 file using spark with the goal that presto can be used later to query from parquet. Basically, there is how it looks like,
Kafka-->Spark-->Parquet<--Presto
I am able to generate parquet in S3 using Spark and its working fine. Now, I am looking at presto and what I think I found is that it needs hive meta store to query from parquet. I could not make presto read my parquet files even though parquet saves the schema. So, does it mean at the time of creating the parquet files, the spark job has to also store metadata in hive meta store?
If that is the case, can someone help me find an example of how it's done. To add to the problem, my data schema is changing, so to handle it, I am creating a programmatic schema in spark job and applying it while creating parquet files. And, if I am creating the schema in hive metastore, it needs to be done keeping this in consideration.
Or could you shed light on it if there is any better alternative way?
You keep the Parquet files on S3. Presto's S3 capability is a subcomponent of the Hive connector. As you said, you can let Spark define tables in Spark or you can use Presto for that, e.g.
create table hive.default.xxx (<columns>)
with (format = 'parquet', external_location = 's3://s3-bucket/path/to/table/dir');
(Depending on Hive metastore version and its configuration, you might need to use s3a instead of s3.)
Technically, it should be possible to create a connector that infers tables' schemata from Parquet headers, but I'm not aware of an existing one.