Post
/

Let's try: dbt part 8 - DAGs and Docs

See how dbt manages the flow and docs for us

data, data engineering
Posted
5 min read
<a href="https://github.com/dbt-labs/dbt-core">dbt-labs/dbt-core</a>
dbt-labs/dbt-core
Let's try: dbt part 8 - DAGs and Docs
In this series
  1. Let's try: dbt part 1 - prologue
  2. Let's try: dbt part 2 - first start
  3. Let's try: dbt part 3 - seed and source
  4. Let's try: dbt part 4 - model materialization
  5. Let's try: dbt part 5 - from Jinja to macros and hooks
  6. Let's try: dbt part 6 - snapshots and analyses
  7. Let's try: dbt part 7 - tests
  8. Let's try: dbt part 8 - DAGs and Docs

I believe that documentation is also a thing. We developed something but if no one could understand it even after reading your papers then it isn’t a good product.

In this blog, we’re gonna see how dbt manages the flow and docs for us.

Model listing

DAG stands for Directed Acyclic Graph. We make models like a river. First model is referred by the second and so on.

dbt can show us the models we created with the command.

1

dbt ls

This command will list all models we have.

For example, we have models in this kind of relationship:

---
title: model lineage
---
flowchart LR
    grade["[source]<br>student_grades"]
    stem["[model]<br>stem_students"]
    lowstem["[model]<br>low_grade_stem_students"]
    highstem["[model]<br>high_grade_stem_students"]
    nonstem["[model]<br>non_stem_students"]
    rep["[model]<br>rep_students"]

    grade --> stem --> lowstem
    stem --> highstem --> rep
    grade --> nonstem --> rep

We should see the list result as follows:

1
2
3
4
5
6
7
8

$ dbt ls

bbz_dbt_test_project.students.high_grade_stem_students
bbz_dbt_test_project.students.low_grade_stem_students
bbz_dbt_test_project.students.non_stem_students
bbz_dbt_test_project.students.rep_students
bbz_dbt_test_project.students.stem_students
source:bbz_dbt_test_project.students.student_grades

Model selection

There are some flags to select models we need to command at a time.

Filter

In dbt, we can simply select or exclude particular models using flag --select and/or --exclude when we execute commands such as ls, build, run, test, etc.

1
2
3
4
5
6

# select
dbt ls --select <model name, path, state, tag, package, etc.>
dbt ls -s <model name, path, state, tag, package, etc.>

# exclude
dbt ls --exclude <model name, path, state, tag, package, etc.>

Graph operator

We may make use of + as a graph operator to define numbers of ancestors (parents of the model) and descendants (children of the model) and @ for selecting every models related to a selected model.

For example:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29

# select a model and 1-level ancestors
dbt ls -s 1+model

# select a model and 3-level ancestors
dbt ls -s 3+model

# select a model and all level ancestors
dbt ls -s +model

# select a model and 1-level descendants
dbt ls -s model+1 

# select a model and 3-level descendants
dbt ls -s model+3

# select a model and all level descendants
dbt ls -s model+

# select a model and 1-level ancestors plus 1-level descendants
dbt ls -s 1+model+1

# select a model and 3-level ancestors plus 3-level descendants
dbt ls -s 3+model+3

# select a model and all level ancestors plus all level descendants
dbt ls -s +model+

# select a model and all level ancestors plus all level descendants plus all level dependencies
dbt ls -s @model

Some examples

Single model
1
2
3
4

$ dbt ls -s high_grade_stem_students

bbz_dbt_test_project.students.high_grade_stem_students


flowchart LR
  grade["[source]<br>student_grades"]
  stem["[model]<br>stem_students"]
  lowstem["[model]<br>low_grade_stem_students"]
  highstem["[model]<br>high_grade_stem_students"]
  nonstem["[model]<br>non_stem_students"]
  rep["[model]<br>rep_students"]

  grade --> stem --> lowstem
  stem --> highstem --> rep
  grade --> nonstem --> rep

  classDef hl stroke:orange,stroke-width:3px;
  highstem:::hl
Multiple models
1
2
3
4
5

$ dbt ls -s high_grade_stem_students non_stem_students

bbz_dbt_test_project.students.high_grade_stem_students
bbz_dbt_test_project.students.non_stem_students


flowchart LR
  grade["[source]<br>student_grades"]
  stem["[model]<br>stem_students"]
  lowstem["[model]<br>low_grade_stem_students"]
  highstem["[model]<br>high_grade_stem_students"]
  nonstem["[model]<br>non_stem_students"]
  rep["[model]<br>rep_students"]

  grade --> stem --> lowstem
  stem --> highstem --> rep
  grade --> nonstem --> rep

  classDef hl stroke:orange,stroke-width:3px;
  highstem:::hl
  nonstem:::hl
Model with ancestors
1
2
3
4
5
6

$ dbt ls -s 2+high_grade_stem_students
  
bbz_dbt_test_project.students.high_grade_stem_students
bbz_dbt_test_project.students.stem_students
source:bbz_dbt_test_project.students.student_grades


flowchart LR
  grade["[source]<br>student_grades"]
  stem["[model]<br>stem_students"]
  lowstem["[model]<br>low_grade_stem_students"]
  highstem["[model]<br>high_grade_stem_students"]
  nonstem["[model]<br>non_stem_students"]
  rep["[model]<br>rep_students"]

  grade --> stem --> lowstem
  stem --> highstem --> rep
  grade --> nonstem --> rep

classDef hl stroke:orange,stroke-width:3px;
  highstem:::hl
  stem:::hl
  grade:::hl
Model with descendants
1
2
3
4
5

$ dbt ls -s high_grade_stem_students+2
  
bbz_dbt_test_project.students.high_grade_stem_students
bbz_dbt_test_project.students.rep_students


flowchart LR
  grade["[source]<br>student_grades"]
  stem["[model]<br>stem_students"]
  lowstem["[model]<br>low_grade_stem_students"]
  highstem["[model]<br>high_grade_stem_students"]
  nonstem["[model]<br>non_stem_students"]
  rep["[model]<br>rep_students"]

  grade --> stem --> lowstem
  stem --> highstem --> rep
  grade --> nonstem --> rep

  classDef hl stroke:orange,stroke-width:3px;
  highstem:::hl
rep:::hl
Model with both ancestors and descendants
1
2
3
4
5
6
7

$ dbt ls -s +high_grade_stem_students+
  
bbz_dbt_test_project.students.high_grade_stem_students
bbz_dbt_test_project.students.rep_students
bbz_dbt_test_project.students.stem_students
source:bbz_dbt_test_project.students.student_grades


flowchart LR
  grade["[source]<br>student_grades"]
  stem["[model]<br>stem_students"]
  lowstem["[model]<br>low_grade_stem_students"]
  highstem["[model]<br>high_grade_stem_students"]
  nonstem["[model]<br>non_stem_students"]
  rep["[model]<br>rep_students"]

  grade --> stem --> lowstem
  stem --> highstem --> rep
  grade --> nonstem --> rep

  classDef hl stroke:orange,stroke-width:3px;
  highstem:::hl
  rep:::hl
stem:::hl
  grade:::hl
Model with all ancestors, descendants, and dependencies
1
2
3
4
5
6
7
8

$ dbt ls -s @high_grade_stem_students
  
bbz_dbt_test_project.students.high_grade_stem_students
bbz_dbt_test_project.students.non_stem_students
bbz_dbt_test_project.students.rep_students
bbz_dbt_test_project.students.stem_students
source:bbz_dbt_test_project.students.student_grades


flowchart LR
  grade["[source]<br>student_grades"]
  stem["[model]<br>stem_students"]
  lowstem["[model]<br>low_grade_stem_students"]
  highstem["[model]<br>high_grade_stem_students"]
  nonstem["[model]<br>non_stem_students"]
  rep["[model]<br>rep_students"]

  grade --> stem --> lowstem
  stem --> highstem --> rep
  grade --> nonstem --> rep

  classDef hl stroke:orange,stroke-width:3px;
highstem:::hl
  nonstem:::hl
  rep:::hl
  stem:::hl
  grade:::hl

non_stem_students has been included because it’s a dependency of rep_students which is a descendant of the selected model.

Model with all ancestors, descendants, dependencies but exclude itself
1
2
3
4
5
6
7

$ dbt ls -s @high_grade_stem_students --exclude high_grade_stem_students
  
bbz_dbt_test_project.students.non_stem_students
bbz_dbt_test_project.students.rep_students
bbz_dbt_test_project.students.stem_students
source:bbz_dbt_test_project.students.student_grades


flowchart LR
  grade["[source]<br>student_grades"]
  stem["[model]<br>stem_students"]
  lowstem["[model]<br>low_grade_stem_students"]
  highstem["[model]<br>high_grade_stem_students"]
  nonstem["[model]<br>non_stem_students"]
  rep["[model]<br>rep_students"]

grade --> stem --> lowstem
  stem --> highstem --> rep
  grade --> nonstem --> rep

  classDef hl stroke:orange,stroke-width:3px;
  nonstem:::hl
  rep:::hl
  stem:::hl
  grade:::hl

Docs

We can see how to list and select models. The thing is dbt has a feature to generate documentation to understand models so easily.

There are 2 commands for docs in dbt.

1
2

dbt docs generate
dbt docs serve

And after waiting for few seconds, the docs should be popped up in a browser like this.

docs

Here we can play around and view the lineage of each model.

Wrap up

  • dbt ls to list models
  • --select to select models
  • --exclude to exclude models
  • selection flags can be use with ls, compile, run, build, test, etc.
  • + to extend the model selection
  • @ to select whole relationship from the selected models, including all ancestors, descendants, and dependencies.
  • dbt docs generate and dbt docs serve to create and view docs.

References

data, data engineering
let's try Python dbt documentation
This post is licensed under CC BY 4.0 by the author.