Why data sharing eliminates data copies in Snowflake - Visual Breakdown

Process Flow - Why data sharing eliminates data copies

Data Provider Creates Share

↓

Share Contains Reference to Data

↓

Consumer Accesses Shared Data

↓

No Data Copy Made

↓

Consumer Queries Data Directly

↓

Data Provider Maintains Single Source

Data sharing works by creating a reference to the original data, so consumers access the same data without making copies.

Execution Sample

Snowflake

CREATE SHARE my_share;
GRANT USAGE ON DATABASE my_db TO SHARE my_share;
-- Consumer: SHOW SHARES LIKE 'my_share';

This code creates a share and grants access to a database without copying data.

Process Table

Step	Action	Data Copy Created?	Result
1	Provider creates share object	No	Share references original data
2	Provider grants USAGE on database to share	No	Share can access data
3	Consumer uses the share	No	Consumer sees data via reference
4	Consumer queries data	No	Data read directly from provider's storage
5	Provider updates data	No	Consumer sees updated data immediately
6	Consumer disconnects	No	No data copies remain

💡 No data copies are made because sharing uses references, not duplication.

Status Tracker

Variable	Start	After Step 1	After Step 2	After Step 3	After Step 4	Final
Share Object	None	Created	Granted Access	Used by Consumer	Active Reference	Exists
Data Copy	None	None	None	None	None	None
Consumer Access	None	None	None	Connected	Querying	Disconnected

Key Moments - 3 Insights

Why doesn't the consumer get a separate copy of the data?

Because the share only contains a reference to the provider's data, no duplication happens (see execution_table steps 1-4).

What happens if the provider updates the data after sharing?

The consumer sees the updated data immediately since they access the same source (see execution_table step 5).

Does disconnecting the consumer remove the shared data?

No, the share object and data remain with the provider; only the consumer's access ends (see execution_table step 6).

Visual Quiz - 3 Questions

Look at the execution_table, at which step does the consumer start querying data?

A Step 4

B Step 2

C Step 3

D Step 5

💡 Hint

Check the 'Action' and 'Result' columns in execution_table rows for when querying starts.

According to variable_tracker, what is the state of 'Data Copy' after step 4?

A Pending

B Created

C None

D Deleted

💡 Hint

Look at the 'Data Copy' row under 'After Step 4' in variable_tracker.

If the provider updates data, what does the consumer see according to the execution_table?

A Old data until consumer reconnects

B Updated data immediately

C No data until share is recreated

D A copy of updated data

💡 Hint

Refer to execution_table step 5 'Result' column.

Concept Snapshot

Data sharing creates a reference, not a copy.
Provider creates a share and grants access.
Consumer queries data directly from provider storage.
Updates by provider reflect immediately.
No extra storage used for copies.

Full Transcript

Data sharing in Snowflake works by creating a share object that references existing data. The provider creates the share and grants USAGE on the database. The consumer then uses this share to query data directly without any data copy being made. This means the consumer always sees the latest data as maintained by the provider. When the consumer disconnects, no copies remain because none were created. This efficient method saves storage and keeps data consistent.