# Auto Grouping Tools
This package comes with a set of helpful auto grouping tools.
These tools solve a problem where you have M:N relations between two entities and need to join them together. Their functionality is also helpful when working with SQL views.
```mysql
SELECT name, surname, job.name as job___name
FROM person
JOIN
works ON works.person_id = person.id
job ON job.id = works.job_id
WHERE
person.id = 2;
```
For a single person, who has two jobs, DB might output something like this:
| name | surname | job___name |
| ----------- | -------------- | ---------- |
| Jane | Doe | Accountant |
| Jane | Doe | Developer |
Which might be OK when fetching only one person and his jobs. There are use cases, when you need to fetch more and more people. Output will be much larger. This is the place, where auto grouping tools come handy.
## auto_group_dict
This function groups dict keys with same prefix under one dict key. Groups used as group keys are identified by group separator `___`.
```python
person = {
"name": "Jane",
"surname": "Doe",
"job___name": "Accountant",
"job___established": 2001
}
ret = auto_group_dict(person)
# Returns in
ret = {
"name": "Jane",
"surname": "Doe",
"job": {
"name": "Accountant",
"established": 2001,
}
}
```
## auto_group_list
__IMPORTANT__: All items which are inside lists are sorted exactly the same as they came from the DB.
Let's say that we have want to retrieve a new person from our DB. Jane Doe now has two jobs: an accountant and a developer.
Database returns two rows as specified above. But in object oriented world, it would be better for us to have it in one dict. This is where `auto_group_list` comes handy.
```sql
SELECT person.name, person.surname, job.name as jobs__name
FROM person
JOIN
works ON works.person_id = person.id
job ON job.id = works.job_id
WHERE
person.id = 2;
```
Assuming our SQL query returns 2 rows like this:
```python
rows = [
{
"name": "Jane",
"surname": "Doe",
"jobs__name": "Accountant",
},
{
"name": "Jane",
"surname": "Doe",
"jobs__name": "Developer",
}
]
ret = auto_group_list(rows)
# Returns in
ret = {
"name": "Jane",
"surname": "Doe",
"jobs": [
{
"name": "Accountant"
},
{
"name": "Developer"
}
]
}
```
This is kind of handy, isn't it? But what if we want to omit our WHERE statement? This is where `auto_group_list_by_pkeys` comes in place.
## auto_group_list_multi
Is enhanced method based on `auto_group_list` functionality. It is best to group more lines of more persons as was mentioned in example above.
Lets imagine the situation you select two or more persons from database with their jobs:
```sql
SELECT person.name, person.surname, job.name as jobs__name
FROM person
JOIN
works ON works.person_id = person.id
job ON job.id = works.job_id
WHERE
person.id in (2, 3)
ORDER BY person.id;
```
Assuming our SQL query returns 4 rows like this:
```python
rows = [
{
"name": "Jane",
"surname": "Doe",
"jobs__name": "Accountant",
},
{
"name": "Jane",
"surname": "Doe",
"jobs__name": "Developer",
},
{
"name": "Jonh",
"surname": "Doesnt",
"jobs__name": "Store manager",
},
{
"name": "John",
"surname": "Doesnt",
"jobs__name": "Destroyer",
}
]
ret = auto_group_list_multi(rows)
# Returns in
ret = [
{
"name": "Jane",
"surname": "Doe",
"jobs": [
{
"name": "Accountant"
},
{
"name": "Developer"
}
],
},
{
"name": "John",
"surname": "Doesnt",
"jobs": [
{
"name": "Store manager"
},
{
"name": "Destoyer"
}
],
}
]
```
How it works? It watchs column values of non-double underscored attributes. If values are changed it groups
previous values of double underscored keys into list.
## auto_group_list_by_pkeys
Next and the last useful is handy when you want to for example fetch multiple people from DB, keep m..n relations and have everything grouped nicely. Like so:
```sql
SELECT person.id as _id, person.name, person.surname, job.name as jobs__name
FROM person
JOIN
works ON works.person_id = person.id
job ON job.id = works.job_id
WHERE
person.id IN (2, 3);
```
Our person no. 2 is Jane Doe, who works as an accountant and a developer. Person no. 3 is John Doe, works as an DevOps Engineer and a developer.
Let's say our grouping key is `_id`.
Our fetched data converted to python might look something like this:
```python
rows = [
{
"_id": 2,
"name": "Jane",
"surname": "Doe",
"jobs__name": "Accountant"
},
{
"_id": 2,
"name": "Jane",
"surname": "Doe",
"jobs__name": "Developer"
},
{
"_id": 3,
"name": "John",
"surname": "Doe",
"jobs__name": "DevOps Engineer"
},
{
"_id": 3,
"name": "John",
"surname": "Doe",
"jobs__name": "Developer"
}
]
```
Let's make it prettier!
```python
ret = auto_group_list_by_pkeys(("_id",), rows, use_auto_group_dict=True)
# Returns dict with 2 items, grouped by key "_id"
ret = {
"2": {
"_id": 2,
"name": "Jane",
"surname": "Doe",
"jobs": [
{
"name": "Accountant"
},
{
"name": "Developer"
}
]
},
"3": {
"_id": 3,
"name": "John",
"surname": "Doe",
"jobs": [
{
"name": "DevOps Engineer"
},
{
"name": "Developer"
}
]
}
}
```
Now we have all our cases covered, ready to go.
## Tools
A few tool functions are avaliable in module `tools`.
List of avaliable functions:
- `dicts_into_list`: Converts dict of dicts into list structure.
- `sort_list_of_dicts`: Sort list of dicts by selected column.
- `dict_pass`: Whitelist dictionary attributes.
- `dict_filter`: Blacklist dictionary attributes.
- `list_into_dict`: Converts list into dict where dict keys are list indexes.
- `dict_swap`: Swap dict keys into values and values into keys.
- `map_dict_into_list`: Maps dict keys into list positions based on keys_map defined in key map list