topic Re: parsing json string value column into dataframe structure in Data Engineering

parsing json string value column into dataframe structure

dipali_globant — Wed, 13 Nov 2024 05:22:16 GMT

Hi All,

I have to read kafka payload which has value column with json string. But the format of the json is as below.

{ "data": [ { "p_al4": "N/A", "p_a5": "N/A", "p_ad": "OA003", "p_aName": "Abc", "p_aFlag": true ,....(dynamic)} ] }

In data key it can have dynamic number of key value pairs.

here "p_" is a column and respective value is given.
Please suggest how to create this payload into dataframe structure.
like below:

+-----+-----+-----+-------+------+------------------+
|p_al4 | p_a5 | p_ad | p_aName | p_aFlag|
+-----+-----+-----+-------+------+----------------+
|N/A | N/A | OU13| Admin | true |
+-----+-----+-----+-------+------+--------------

Re: parsing json string value column into dataframe structure

SparkJun — Wed, 13 Nov 2024 05:58:34 GMT

Would you know the element in the Kafka event JSON ahead of the time or count the number of element each time and then generate the schema for such and pass it from_json to parse? https://spark.apache.org/docs/latest/api/python/reference/pyspark.sql/api/pyspark.sql.functions.from_json.html

Re: parsing json string value column into dataframe structure

dipali_globant — Wed, 13 Nov 2024 06:30:11 GMT

No I don't know element in JSON . so I can't define structure.