I have created a pyspark dataframe and trying to write the file in s3 bucket in csv format. here the file is writing in csv but the issue is it's writing the file twice(i.e., with actual data and another is with empty data). I have checked the data frame by printing fine only. please suggest any way to prevent that empty wouldn't create.
code snippet:
df = spark.createDataFrame(data=dt1, schema = op_df.columns)
df.write.option("header","true").csv("s3://"+ src_bucket_name+"/src/output/"+row.brand +'/'+fileN)
One possible solution to make sure that the output will include only one file is to do repartition(1) or coalesce(1) before writing.
So something like this:
df.repartition(1).write.option("header","true").csv("s3://"+ src_bucket_name+"/src/output/"+row.brand +'/'+fileN)
Note that having one partition doesn't not necessarily mean that it will result in one file as this can depend on the spark.sql.files.maxRecordsPerFile configuration as well. Assuming this config is set to 0 (the default) you should get only 1 file in the output.
Related
I have a list of DataFrames by the name of "dfs" which has 139 DataFrames. I originally imported csv files to python and have deleted first few rows from each data frame. Now I wish to save these new files back in their original positions. How can I do that. My new data is saved in another list named final. Also please tell me if I can make my code more efficient as I am a new to python.
dfs = [pd.read_csv(filename) for filename in filenames]
final=[]
for i in range(139):
a= dfs[i].iloc[604:,]
final.append(a)
Not sure if I've understood it correctly, if you want to write df to csv to the same as when you made df but this time opposite way.
for df, filename in zip(final, filenames):
path = f'{filename}.csv'
df.to_csv(path)
I have a pyspark job which writes my resultant dataframe in local filesystem. Currently it is running in local mode and so I am doing coalesce(1) to get a single file as below
file_format = 'avro' # will be dynamic and so it will be like avro, json, csv, etc
df.coalesce.write.format(file_format).save('file:///pyspark_data/output')
But I see a lot of memory issues (OOM) and takes longer time as well. So I want to run this job with master as yarn and mode as client. And so to write the result df into a single file in localsystem, I need to use toLocalIterator which yields Rows. How can I stream these Rows into a file of required format (json/avro/csv/parquet and so on)?
file_format = 'avro'
for row in df.toLocalIterator():
# write the data into a single file
pass
You get OOM error because you try to retrieve all the data into a single partition with: coalesce(1)
I dont recommend to use toLocalIterator because you will re-rewrite a custom writer for every format and you wont have parallele writing.
You first solution is a good one :
df.write.format(file_format).save('file:///pyspark_data/output')
if you use hadoop you can retrieve all the data into one on filesysteme this way : (it work for csv, you can try for other) :
hadoop fs -getmerge <HDFS src> <FS destination>
i have a few txt files in a directory(i have only the path and not the names of the files) that contain json data,and i need to read all of them into a dataframe.
i tried this:
df=sc.wholeTextFiles("path/*")
but i cant even display the data and my main goal is to preform queries in diffrent ways on the data.
Instead of wholeTextFiles(gives key, value pair having key as filename and data as value),
Try with read.json and give your directory name spark will read all the files in the directory into dataframe.
df=spark.read.json("<directorty_path>/*")
df.show()
From docs:
wholeTextFiles(path, minPartitions=None, use_unicode=True)
Read a directory of text files from HDFS, a local file system
(available on all nodes), or any Hadoop-supported file system URI.
Each file is read as a single record and returned in a key-value pair,
where the key is the path of each file, the value is the content of
each file.
Note: Small files are preferred, as each file will be loaded fully in
memory.
I have written a scraper which downloads daily flight prices, stores them as pandas data frames and saves them off as csv files in a given folder. I am now trying to combine these csv files into pandas for data analysis using append, but end result is an empty data frame.
Specifically, individual csv files are loaded correctly into pandas, but the append seems to fail (and several methods found on stackoverflow posts don't seem to work). Code is below, any pointers? Thanks!
directory = os.path.join("C:\\Testfolder\\")
for root,dirs,files in os.walk(directory):
for file in files:
daily_flight_df = (pd.read_csv(directory+file,sep=";")) #loads csv into dataframe - works correctly
cons_flight_df.append(daily_flight_df) #appends daily flight prices into a pandas with consolidated flight prices - does not seem to work
print(cons_flight_df) #currently prints out an empty data frame
cons_flight_df.to_csv('C:\\Testfolder\\test.csv') #currently returns empty csv file
In pandas, the append method isn't in place. You need to assign it.
cons_flight_df = cons_flight_df.append(daily_flight_df)
We use Pig to load files from directories containing thousands of files, transform them, and then output files that are a consolidation of the input.
We've noticed that the output files contain the header record of every file processed, i.e. the header appears multiple times in each file.
Is there any way to have the header only once per output file?
raw_data = LOAD '$INPUT'
USING org.apache.pig.piggybank.storage.CSVExcelStorage(',')
DO SOME TRANSFORMS
STORE data INTO '$OUTPUT'
USING org.apache.pig.piggybank.storage.CSVExcelStorage('|')
Did you try this option?
SKIP_INPUT_HEADER
See https://github.com/apache/pig/blob/31278ce56a18f821e9c98c800bef5e11e5396a69/contrib/piggybank/java/src/main/java/org/apache/pig/piggybank/storage/CSVExcelStorage.java#L85