--- title: "Load from Amazon S3" component: "savanna" version: "main" module: "graph-development" html_url: "/savanna/main/graph-development/load-data/load-from-s3.html" --- [View as HTML](/savanna/main/graph-development/load-data/load-from-s3.html) · [Documentation index](/llms.txt) # Load from Amazon S3 If you store your data in Amazon S3, TigerGraph Savanna provides seamless integration for data ingestion. You can directly load data from your S3 buckets into your graph databases, eliminating the need for manual data transfers. This simplifies the process of importing large datasets and enables you to leverage the scalability and durability of Amazon S3 for your graph analysis. ## 1) Select Source Once you've selected **Amazon S3**, you will be asked to configure the Amazon S3 data source. 1. Click on ![Screenshot 2024 04 17 at 9.36.58 PM](../_images/Screenshot%202024-04-17%20at%209.36.58 PM.png) to add a new Amazon S3 data source. ![Screenshot 2024 04 17 at 9.36.27 PM](../_images/Screenshot%202024-04-17%20at%209.36.27 PM.png) 2. You will need to provide your S3 `AWS access key id` and S3 `AWS secret access key`. ![Screenshot 2024 04 17 at 9.37.32 PM](../_images/Screenshot%202024-04-17%20at%209.37.32 PM.png) 3. Once you have those configured, you can add one or multiple `S3 URI` within the same S3 bucket. 4. Click **Next** to process the file(s). > [!NOTE] > The current data loading tool only supports CSV,TSV and JSON files. Other formats will be available in later releases. ## 2) Configure File This step lets you configure the source file details. 1. The data loading tool will automatically detect the `.csv` separators and line breaks. The parser automatically splits each line into a series of tokens. ![config file](../_images/config-file.png) > [!NOTE] > If the parsing is **not** correct, click on the ![Screenshot 2024 04 17 at 5.54.17 PM](../_images/Screenshot%202024-04-17%20at%205.54.17 PM.png)button to configure a different option for the delimiter, such as `eol`, or `quote and header`.![Screenshot 2024 04 17 at 5.54.50 PM](../_images/Screenshot%202024-04-17%20at%205.54.50 PM.png) The enclosing character is used to mark the boundaries of a token, overriding the delimiter character. For example, if your delimiter is a comma, but you have commas in some strings, then you can define single or double quotes as the enclosing character to mark the endpoints of your string tokens. > [!NOTE] > It is not necessary for every token to have enclosing characters. The parser uses enclosing characters when it encounters them. > [!NOTE] > You can edit the header line of the parsing result to give each column a more intuitive name, since you will be referring to these names when loading data to the graph. The header name is ignored during data loading. 2. Once you are satisfied with the file settings, click **Next** to proceed. ## 3) Configure Map If you are loading data into a brand new graph, you will be prompted to let our engine generate a schema and mapping for you. Or you can start from scratch. For more details of schema design please refer to [Design Schema](../design-schema/index.md). 1. Select `Generate the schema only` or `Generate the schema and data mapping`. ![Screenshot 2024 04 17 at 5.55.35 PM](../_images/Screenshot%202024-04-17%20at%205.55.35 PM.png) > [!NOTE] > The schema generation feature is still a preview feature. The correctness and efficiency of the resulting graph schema and mapping could vary. 2. In the `Source` column, you can choose the specific column from the data source that you want to map with the attribute. ![config mapping 1](../_images/config-mapping-1.png) 3. Use the `+` button to create a new attribute of the target vertex or edge. ![config mapping 2](../_images/config-mapping-2.png) 4. Click the **Token Function** button to configure token functions for the selected source. For more details of configuring token functions, please refer to [Token Function](token-function.md). 5. Click the **Quick Map** button to quickly map the data source headers to the existing schema attributes. ![quick map](../_images/quick-map.png) 1. The **Map all to target** button aligns existing attribute names with the corresponding data source headers, it won't introduce new attributes. 2. The **Map all from source** button not only aligns existing attribute names with the corresponding data asource headers, but also introduces new attributes based on unmatched data source headers. 3. The following list shows the mapping status of each attribute. You can manually adjust the mapping by checking the box next to the attribute name. 6. Click **Next** to proceed. ## 4) Confirm This step will let you confirm the changes made to the schema and the data mapping you created to load the data. 1. Simply review the `Schema to be changed` and `Data to be loaded` lists. ![confirm](../_images/confirm.png) > [!NOTE] > Please be aware that some schema changes will result in unintentional deletion of the data. Please carefully review the warning message before confirming the loading. 2. Click on the **Confirm** button to run the loading jobs and monitor their `Status`. ![Screenshot 2024 04 17 at 5.59.16 PM](../_images/Screenshot%202024-04-17%20at%205.59.16 PM.png) ## Next Steps Next, learn how to use [Design Schema](../design-schema/index.md), [GSQL Editor](../gsql-editor/index.md) and [Explore Graph](../explore-graph/index.md) in TigerGraph Savanna. Or return to the [Overview](../../overview/index.md) page for a different topic.