# Non-historized or Sat Dependent Child Key

**URL:** <https://forum.data-community.org/t/non-historized-or-sat-dependent-child-key/1160>\
**Category:** Data Vault 2.0\
**Created:** [February 7, 2024, 11:39pm UTC](https://forum.data-community.org/t/non-historized-or-sat-dependent-child-key/1160 "2024-02-07T23:39:04Z")\
**Posts on this page:** 7\
**Page:** 1

<div class="post-metadata">

**Author:** ![rudra](https://avatars.discourse-cdn.com/v4/letter/r/47e85d/32.png) [@rudra](https://forum.data-community.org/u/rudra)\
**Post date:** [February 7, 2024, 11:39pm UTC](https://forum.data-community.org/t/non-historized-or-sat-dependent-child-key/1160/1 "2024-02-07T23:39:04Z")

</div>

Hello,

I need to design for Orders, Order Lines, Shipment. In summary, they all from sales order process.

These table contains main key details such as  
**Orders** :  
Order number  
Customer Number  
Order Type, Order Date, etc

```
**Order Line:**
    Order number 
    Order Line Number
    Customer Number 
    Item Number

```

I am inclined to implement Sat Child Dependent Child key, rather than non-historized as this is good candidate for streaming use case or single business concept - Hub.

 ![image](https://europe1.discourse-cdn.com/flex013/uploads/datavaultusergroup/original/1X/e43530f1319b93fb5dfa49db48f156225edaf6c5.png)

In addition to above, should i also create a Link\_SALES\_ORDER\_CUSTOMER\_ITEM to connect with Hub\_Customer and Hub\_Item? Link table ensure unique relationship across hub’s, then what are the implications on the dependent child key - Order Line Number in the exmaple?

Can you please share your thoughts / ideas!

Thank you.

---

<div class="post-metadata">

**Author:** ![patrickcuba](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.data-community.org/patrickcuba/32/19_2.png) [@patrickcuba](https://forum.data-community.org/u/patrickcuba)\
**Post date:** [February 8, 2024, 2:23am UTC](https://forum.data-community.org/t/non-historized-or-sat-dependent-child-key/1160/2 "2024-02-08T02:23:20Z")

</div>

non-historised = streaming / near-real time ingestion; kind of why it was renamed from t-links to emphasize that  
link-sat with a dep-key = batch ingestion works best, it’s idempotent

---

<div class="post-metadata">

**Author:** ![rudra](https://avatars.discourse-cdn.com/v4/letter/r/47e85d/32.png) [@rudra](https://forum.data-community.org/u/rudra)\
**Post date:** [February 8, 2024, 3:19am UTC](https://forum.data-community.org/t/non-historized-or-sat-dependent-child-key/1160/3 "2024-02-08T03:19:16Z")

</div>

Thanks @patrickcuba . Are you saying create a Link table with link-sat with a dep-key like below?

 ![image](https://europe1.discourse-cdn.com/flex013/uploads/datavaultusergroup/original/1X/c6a8a9c0c102cb7e9968cfb7906d2a39d5bf16a6.png)

Dep-Key - Order Line number - should be part of Lnk\_HK calculation?

---

<div class="post-metadata">

**Author:** ![patrickcuba](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.data-community.org/patrickcuba/32/19_2.png) [@patrickcuba](https://forum.data-community.org/u/patrickcuba)\
**Post date:** [February 8, 2024, 3:37am UTC](https://forum.data-community.org/t/non-historized-or-sat-dependent-child-key/1160/4 "2024-02-08T03:37:38Z")

</div>

I generally don’t include dep-keys in link-hash calc, because it isn’t a business/primary object. It also means I can reuse that same relationship for other pipelines to load to the same link table and thus have the same link-hash-key.

Order line makes a good dep-key, it needs the order to live!

---

<div class="post-metadata">

**Author:** ![rudra](https://avatars.discourse-cdn.com/v4/letter/r/47e85d/32.png) [@rudra](https://forum.data-community.org/u/rudra)\
**Post date:** [February 8, 2024, 3:44am UTC](https://forum.data-community.org/t/non-historized-or-sat-dependent-child-key/1160/5 "2024-02-08T03:44:08Z")

</div>

So this means exclude Order Line from link-hash-key, rather keep order line as dep-key in Sat table, and this process will be idempotent.

Hope I’m following you correctly!

---

<div class="post-metadata">

**Author:** ![patrickcuba](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.data-community.org/patrickcuba/32/19_2.png) [@patrickcuba](https://forum.data-community.org/u/patrickcuba)\
**Post date:** [February 9, 2024, 12:25am UTC](https://forum.data-community.org/t/non-historized-or-sat-dependent-child-key/1160/6 "2024-02-09T00:25:45Z")

</div>

> [@rudra](#):
>
> this process will be idempotent.

Yes, if you execute the load twice the outcome is still the same

---

<div class="post-metadata">

**Author:** ![rudra](https://avatars.discourse-cdn.com/v4/letter/r/47e85d/32.png) [@rudra](https://forum.data-community.org/u/rudra)\
**Post date:** [February 9, 2024, 4:30pm UTC](https://forum.data-community.org/t/non-historized-or-sat-dependent-child-key/1160/7 "2024-02-09T16:30:03Z")

</div>

Thank you so much for your inputs @patrickcuba
