File formats (CSV, JSON, Parquet, Avro) in Snowflake - Step-by-Step Execution

Process Flow - File formats (CSV, JSON, Parquet, Avro)

Start: Data to store

↓

Choose file format

↓

Data saved in chosen format

↓

Data loaded by Snowflake

↓

Data parsed and used in queries

↓

End

This flow shows how data is saved in different file formats and then loaded and used in Snowflake.

Execution Sample

Snowflake

CREATE OR REPLACE FILE FORMAT my_csv_format
  TYPE = 'CSV'
  FIELD_DELIMITER = ','
  SKIP_HEADER = 1;

COPY INTO my_table FROM @my_stage FILE_FORMAT = (FORMAT_NAME = 'my_csv_format');

This code creates a CSV file format and loads data from a stage into a table using that format.

Process Table

Step	Action	File Format	Effect	Snowflake Behavior
1	Create CSV file format	CSV	Defines delimiter and header skip	Snowflake knows how to parse CSV files
2	Load data from stage	CSV	Reads CSV files with defined format	Data is parsed into table rows
3	Query data	CSV	Uses parsed data	Returns table rows for queries
4	Create JSON file format	JSON	Defines JSON parsing rules	Snowflake parses JSON objects
5	Load JSON data	JSON	Reads JSON files	Data loaded as semi-structured data
6	Create Parquet file format	Parquet	Defines columnar binary format	Snowflake reads efficient columnar data
7	Load Parquet data	Parquet	Reads Parquet files	Data loaded with schema and compression
8	Create Avro file format	Avro	Defines schema-based binary format	Snowflake reads Avro files with schema
9	Load Avro data	Avro	Reads Avro files	Data loaded with schema validation
10	Query all data	All	Uses internal parsing	Snowflake returns data rows correctly
11	End	-	-	Data ready for use in Snowflake

💡 All data formats are parsed and loaded successfully for querying in Snowflake.

Status Tracker

Variable	Start	After CSV Load	After JSON Load	After Parquet Load	After Avro Load	Final
Data in Stage	Raw files	CSV files	JSON files	Parquet files	Avro files	All files ready
File Format Object	None	CSV format defined	JSON format defined	Parquet format defined	Avro format defined	All formats defined
Table Data	Empty	CSV data loaded	JSON data loaded	Parquet data loaded	Avro data loaded	All data loaded

Key Moments - 3 Insights

Why do we need to define a file format before loading data?

Snowflake needs to know how to read the file correctly. For example, CSV files need delimiter info, JSON needs parsing rules. See execution_table rows 1 and 4.

What happens if the file format does not match the actual file?

Snowflake will fail to parse the data correctly, causing errors or wrong data. This is why defining the correct format is important (rows 2, 5, 7, 9).

How does Snowflake handle different file formats internally?

Snowflake uses specific parsers for each format to convert files into table rows for queries, as shown in rows 3, 5, 7, 9, and 10.

Visual Quiz - 3 Questions

Look at the execution_table, at which step is the Parquet file format created?

A Step 8

B Step 3

C Step 6

D Step 4

💡 Hint

Check the 'Action' and 'File Format' columns in execution_table rows.

According to variable_tracker, what is the state of 'Table Data' after JSON load?

A Empty

B JSON data loaded

C CSV data loaded

D All data loaded

💡 Hint

Look at the 'Table Data' row and the 'After JSON Load' column in variable_tracker.

If the CSV file format is not defined correctly, what will happen during loading?

A Loading will fail or data will be incorrect

B Data will load but queries will be slow

C Snowflake will guess the format and load data correctly

D Snowflake will convert CSV to JSON automatically

💡 Hint

Refer to key_moments about file format importance and execution_table steps 1 and 2.

Concept Snapshot

File formats in Snowflake:
- Define file format objects (CSV, JSON, Parquet, Avro)
- Specify parsing rules (delimiter, schema, etc.)
- Load data from stage using file format
- Snowflake parses files into table rows
- Correct format ensures successful data load and query

Full Transcript

This visual execution shows how Snowflake handles different file formats: CSV, JSON, Parquet, and Avro. First, you create a file format object that tells Snowflake how to read the files. For example, CSV needs delimiter info, JSON needs parsing rules, Parquet and Avro use schemas. Then, you load data from a stage using the defined format. Snowflake parses the files and loads data into tables. Finally, you can query the data as normal. The execution table traces each step from creating formats to loading and querying data. The variable tracker shows how data and formats change state during execution. Key moments clarify why defining the correct format is crucial and what happens if formats don't match. The quiz tests understanding of steps and states. This helps beginners see how Snowflake processes different file formats step-by-step.