One Lake

Hi Lachu1935 ,

The issue here is that Fabric will not reliably auto-detect ambiguous date types (non ISO dates). As you are importing as a csv it will struggle to identify your dates as dates, read them as a string and format as text.

The solution here is to do one of the following:

1. Structure your dates in your source data to a common ISO format like 2026-01-30 (yyyy‑mm‑dd)
2. Change the column format using a notebook as Cookistador suggests (but this seems to be not working)
3. Bring your data in using Dataflow Gen2 and using the Power Query editor to change the column format

I would personally get into the habit of using ISO date formats in data reporting if possible. You will have far less trouble when ingesting and transforming data. 

EDIT:

Have you tried Data Wrangler in notebooks? This assists you in transforming data within notebooks.

df = spark.read.option("header", "true").csv("Files/myfolder/myfile.csv")

display(df)

Accelerate Data Prep with Data Wrangler - Microsoft Fabric | Microsoft Learn

Hello Lachu1935 

It is best to drop the CSV file in /Files (consider it a Landing zone), read the content using PySpark explicitly specifying the schema, and finally write to /Tables area in your "bronze" schema. I attach a code snippet. 

from pyspark.sql.types import StructType, StructField, StringType, IntegerType, DoubleType, DateType

schema = StructType([
    StructField("order_id",     StringType(), True),
    StructField("customer_id",  StringType(), True),
    StructField("order_date",   DateType(),   True),   # <-- DateType; parsed using dateFormat
    StructField("channel",      StringType(), True),
    StructField("quantity",     IntegerType(),True),
    StructField("unit_price",   DoubleType(), True)
])

src_path = "/lakehouse/default/Files/ingest/orders/"

df = (
    spark.read
         .format("csv")
         .option("header", "true")          # CSV has headers
         .option("mode", "PERMISSIVE")      # don't fail job on a few bad lines
         .option("dateFormat", "dd-MM-yyyy")# <-- critical: tell Spark how to parse DateType
         .schema(schema)
         .load(src_path)
)

spark.sql("CREATE DATABASE IF NOT EXISTS bronze")

(df.write
   .format("delta")
   .mode("append")              # or "overwrite" for full refresh
   # .partitionBy("order_date") # uncomment if partitioning is helpful for your volume
   .saveAsTable("bronze.orders")
)

Hi Lachu1935 

it is probably related to the format of your date

It is waiting date in us format so this is why it is recognized as text (in which area is deployed your fabric capacity?)

Instead of doing a auto load from file to table, you can create a small notebook to convert the column in the right format:

Thank you Team for such a meaningful guidance from you all , the issue was now resloved thanlks for the support Thank you

I have tried this, but the column still changes to all null values. Could this be an issue with the CSV file? I have checked, and the column shows as a data format.