Referencing common table in files with various names

[Kscott10165]

To set the stage, I am working on building a flow where users are able to drop excel files in a sharepoint folder. The flow references one specific table and, for each row in the table, creates a new entry on a list with the information of that table row.

The hangup is being able to reference a specific table when there is no set file name and referencing the columns in that table when creating the new list entry. Any ideas?

[grantjenkins]

A few questions:

1. Is the trigger, When a file is created (properties only)?
2. Will the Table name always be the same, or could it be anything?
3. If not a specific Table name, will each Excel File potentially contain multiple Tables, or just one?
4. Will the columns in the Table always be the same?

[Kscott10165]

1. Is the trigger, When a file is created (properties only)? Yes. 
2. Will the Table name always be the same, or could it be anything? The table name will always be the same. "Employees" the excel file will be a standardized form but they are often renamed.
3. If not a specific Table name, will each Excel File potentially contain multiple Tables, or just one? There will be a few tables on the sheet but only one will need to be referenced. 
4. Will the columns in the Table always be the same? The columns will always be the same as will the column names. There will be a varying number of rows in the table depending on the number of personnel being submitted.

Thanks for the reply! This flow could save my team a lot of time.

[grantjenkins · Accepted solution]

Here's a couple of examples that will hopefully get what you're after.

I have the following SharePoint List.

And I have a folder in my library where the Excel files will be uploaded to. It doesn't matter what the file names are, as long as they have a Table called SalesTable and the appropriate columns. As an example:

EXAMPLE 1 - Using JSON Schema

The full flow is below. I'll go into each of the actions.

When a file is created (properties only) is set to my Dropoff folder.

List rows present in a table uses the Full Path from the trigger and has the Table name hardcoded to SalesTable.

Before moving to the next step, you should save and run the flow, then copy the output from List rows present in a table. We will use this to generate the schema in the next step. Once you've copied the output, go back to edit mode and proceed with the next step.

Parse JSON takes in the value from List rows present in a table. To get the schema click on Generate from sample, then paste in the output you copied from the previous step and click on Done.

In this example, the schema would look like the following:

{
    "type": "array",
    "items": {
        "type": "object",
        "properties": {
            "@@odata.etag": {
                "type": "string"
            },
            "ItemInternalId": {
                "type": "string"
            },
            "Title": {
                "type": "string"
            },
            "First Name": {
                "type": "string"
            },
            "Last Name": {
                "type": "string"
            },
            "Email": {
                "type": "string"
            },
            "Country": {
                "type": "string"
            },
            "Company": {
                "type": "string"
            }
        },
        "required": [
            "@@odata.etag",
            "ItemInternalId",
            "Title",
            "First Name",
            "Last Name",
            "Email",
            "Country",
            "Company"
        ]
    }
}

The @@odata.etag and ItemInternalId fields are auto-generated. You can just ignore them or remove them from the schema. Also, if some of your columns could potentially contain empty data (not all fields filled in) then you would need to remove those fields from the "required" section. For this example, I'll remove the auto-generated fields and assume that Country is optional (not required). My updated schema would look like:

{
    "type": "array",
    "items": {
        "type": "object",
        "properties": {
            "Title": {
                "type": "string"
            },
            "First Name": {
                "type": "string"
            },
            "Last Name": {
                "type": "string"
            },
            "Email": {
                "type": "string"
            },
            "Country": {
                "type": "string"
            },
            "Company": {
                "type": "string"
            }
        },
        "required": [
            "Title",
            "First Name",
            "Last Name",
            "Email",
            "Company"
        ]
    }
}

The Apply to each iterates over each of the rows (using the Body from Parse JSON), and for each one, creates a new item in the list.

And that's it - you should now have items been added into your list.

EXAMPLE 2 - NOT using JSON Schema

As an alternative option, you could bypass the Parse JSON and reference the fields directly. See full flow below:

For the Apply to each, you would now just pass in value from List rows present in a table.

And for each of the values you would use the following expressions:

//Example
items('Apply_to_each')?['Excel_Column_Name']

//Actual
items('Apply_to_each')?['Title']
items('Apply_to_each')?['First Name']
items('Apply_to_each')?['Last Name']
items('Apply_to_each')?['Email']
items('Apply_to_each')?['Country']
items('Apply_to_each')?['Company']

This would give you the same result as option 1 and might be a bit easier for your scenario.

[Kscott10165]

Thank you so much! Works like a dream!