I have a requirement to scan a FixedWidth file using a specific schema and once this is done, the resulted DF with filters applied needs to be converted back to fixed width. How can we apply such transformations before the file being saved to s3. Below is what I have done.
df = spark.read.text(dataset_path)
# Dataframe with applied selection logic
df = df.select(
df.value.substr(1, 10).alias('name'),
df.value.substr(11, 20).alias('another_name'),
df.value.substr(31, 60).alias('address')
)
df = df.filter(df.name.isin('some_name'))
# Here is the dataframe which I need to convert to FixedWidth before saving.
df.save('s3a://somebucket/somepath')
Is there a way to get this done in PySpark?
Related
I'm trying to run real-time inference on a spark structured stream, first I trained the model
#model_creation
model.fit()
model.predict([ 33.26, 68.51, 1012.49, 52.68])
#create spark df from kafka stream
df = spark.readstream.format("kafka").....
#inference
def predict(input):
#extract json from input and conver to list of doubles
#model.predict(input_array)
#result = model.predict(input_list)
#return result
spark.udf.register("lr_predict", predict ,StringType())
df3 = df2.withColumn('predict_response',predict(col('value')))
display(df3)
I'm not sure how to extract the json input from sql spark dataframe and run it in model, I've been trying things since yesterday nothing seem to stick.
import json
def predict(input):
""" I am creating the function with following assumptions
please tell me if those are not correct
the model outputs a string value
the model need numeric data type as input
"""
#extract json from input and conver to list of doubles
#model.predict(input_array)
#result = model.predict(input_list)
#return result
input_proccesed=json.loads(input)
input_features_array=input_proccesed['input']
##casting to float just to be sure
input_features_array=[float(x) for x in input_features_array]
#predicting the output
result= model.predict(input_proccesed['input'])
return result
##if the output is string like - "GOOD" /"BAD" the use StringType()
##if the output is numeric like - 0.92 ,1028.384 the use DecimalType()
from pyspark.sql.types import StringType,DecimalType,FloatType,DoubleType
from pyspark.sql.functions import udf,col
lr_prediction=udf(predict,DoubleType())
df3 = df2.withColumn('predict_response_as_double',lr_prediction(col('value')))
lr_prediction=udf(predict,StringType())
df4 = df2.withColumn('predict_response_as_string',lr_prediction(col('value')))
I am trying to convert my pyspark sql dataframe to json and then save as a file.
df_final = df_final.union(join_df)
df_final contains the value as such:
I tried something like this. But it created a invalid json.
df_final.coalesce(1).write.format('json').save(data_output_file+"createjson.json", overwrite=True)
{"Variable":"Col1","Min":"20","Max":"30"}
{"Variable":"Col2","Min":"25,"Max":"40"}
My expected file should have data as below:
[
{"Variable":"Col1",
"Min":"20",
"Max":"30"},
{"Variable":"Col2",
"Min":"25,
"Max":"40"}]
For pyspark you can directly store your dataframe into json file, there is no need to convert the datafram into json.
df_final.coalesce(1).write.format('json').save('/path/file_name.json')
and still you want to convert your datafram into json then you can used
df_final.toJSON().
A solution can be using collect and then using json.dump:
import json
collected_df = df_final.collect()
with open(data_output_file + 'createjson.json', 'w') as outfile:
json.dump(data, outfile)
Here is how you can do the equivalent of json.dump for a dataframe with PySpark 1.3+.
df_list_of_jsons = df.toJSON().collect()
df_list_of_dicts = [json.loads(x) for x in df_list_of_jsons]
df_json = json.dumps(df_list_of_dicts)
sc.parallelize([df_json]).repartition(1).cache().saveAsTextFile("<HDFS_PATH>")
Note this will result in the whole dataframe being loaded into the driver memory, so this is only recommended for small dataframe.
If you want to use spark to process result as json files, I think that your output schema is right in hdfs.
And I assumed you encountered the issue that you can not smoothly read data from normal python script by using :
with open('data.json') as f:
data = json.load(f)
You should try to read data line by line:
data = []
with open("data.json",'r') as datafile:
for line in datafile:
data.append(json.loads(line))
and you can use pandas to create dataframe :
df = pd.DataFrame(data)
I am creating an empty dataframe and later trying to append another data frame to that. In fact I want to append many dataframes to the initially empty dataframe dynamically depending on number of RDDs coming.
the union() function works fine if I assign the value to another a third dataframe.
val df3=df1.union(df2)
But I want to keep appending to the initial dataframe (empty) I created because I want to store all the RDDs in one dataframe. The below code however does not show right counts. It seems that it simply did not append
df1.union(df2)
df1.count() // this shows 0 although df2 has some data and that is shown if I assign to third datafram.
If I do the below (I get reassignment error since df1 is val. And if I change it to var type, I get kafka multithreading not safe error.
df1=d1.union(df2)
Any idea how to add all the dynamically created dataframes to one initially created data frame?
Not sure if this is what you are looking for!
# Import pyspark functions
from pyspark.sql.types import StructType, StructField, IntegerType, StringType
# Define your schema
field = [StructField("Col1",StringType(), True), StructField("Col2", IntegerType(), True)]
schema = StructType(field)
# Your empty data frame
df = spark.createDataFrame(sc.emptyRDD(), schema)
l = []
for i in range(5):
# Build and append to the list dynamically
l = l + [([str(i), i])]
# Create a temporary data frame similar to your original schema
temp_df = spark.createDataFrame(l, schema)
# Do the union with the original data frame
df = df.union(temp_df)
df.show()
DataFrames and other distributed data structures are immutable, therefore methods which operate on them always return new object. There is no appending, no modification in place, and no ALTER TABLE equivalent.
And if I change it to var type, I get kafka multithreading not safe error.
Without actual code is impossible to give you a definitive answer, but it is unlikely related to union code.
There is a number of known Spark bugs cause by incorrect internal implementation (SPARK-19185, SPARK-23623 to enumerate just a few).
I have a data frame in PySpark called df. I have registered this df as a temptable like below.
df.registerTempTable('mytempTable')
date=datetime.now().strftime('%Y-%m-%d %H:%M:%S')
Now from this temp table I will get certain values, like max_id of a column id
min_id = sqlContext.sql("select nvl(min(id),0) as minval from mytempTable").collect()[0].asDict()['minval']
max_id = sqlContext.sql("select nvl(max(id),0) as maxval from mytempTable").collect()[0].asDict()['maxval']
Now I will collect all these values like below.
test = ("{},{},{}".format(date,min_id,max_id))
I found that test is not a data frame but it is a str string
>>> type(test)
<type 'str'>
Now I want save this test as a file in HDFS. I would also like to append data to the same file in hdfs.
How can I do that using PySpark?
FYI I am using Spark 1.6 and don't have access to Databricks spark-csv package.
Here you go, you'll just need to concat your data with concat_ws and right it as a text:
query = """select concat_ws(',', date, nvl(min(id), 0), nvl(max(id), 0))
from mytempTable"""
sqlContext.sql(query).write("text").mode("append").save("/tmp/fooo")
Or even a better alternative :
from pyspark.sql import functions as f
(sqlContext
.table("myTempTable")
.select(f.concat_ws(",", f.first(f.lit(date)), f.min("id"), f.max("id")))
.coalesce(1)
.write.format("text").mode("append").save("/tmp/fooo"))
I am building a Python script in which I need to generate a json file from json RDD .
Following is code snippet for saving json file.
jsonRDD.map(lambda x :json.loads(x))
.coalesce(1, shuffle=True).saveAsTextFile('examples/src/main/resources/demo.json')
But I need to write json data to a single file instead of data distributed across several partitions.
So please suggest me appropriate solution for it
Without the use of additional libraries like pandas, you could save your RDD of several jsons by reducing them to one big string of jsons, each separated by a new line:
# perform your operation
# note that you do not need a lambda expression for json.loads
jsonRDD = jsonRDD.map(json.loads).coalesce(1, shuffle=True)
# map jsons back to string
jsonRDD = jsonRDD.map(json.dumps)
# reduce to one big string with one json on each line
json_string = jsonRDD.reduce(lambda x, y: x + "\n" + y)
# write your string to a file
with open("path/to/your.json", "w") as f:
f.write(json_string.encode("utf-8"))
I have had issues with pyspark saving off JSON files once I have them in a RDD or dataframe, so what I do is convert them to a pandas dataframe and save them to a non distributed directory.
import pandas
df1 = sqlContext.createDataFrame(yourRDD)
df2 = df1.toPandas()
df2.to_json(yourpath)