Ignore:
Timestamp:
09/29/26 20:55:13 (2 hours ago)
Author:
Stefan <trsunovstefan@…>
Branches:
main
Parents:
0cee8ec
Message:

Correct P1/P2 consistency (Holdings, WatchlistItems), redo P5 normalization

File:
1 edited

Legend:

Unmodified
Added
Removed
  • docs/P2-RelationalDesign/RelationalDesign.md

    r0cee8ec r1549dae  
    11# Relational Design
    22
     3This page transforms [ERModel](../P1-ConceptualModel/ERModel.md) **v05** into
     4relations. Every relation below corresponds to exactly one entity set of the
     5model, and every foreign key corresponds to exactly one relationship, so the
     6two diagrams can be compared box for box and line for line (see
     7[Relational diagram](#relational-diagram)).
     8
    39## Descriptive representation of the relational schema
    410
    5 Notation: **bold** = primary key, *italic* = foreign key.
    6 
    7 - **Users**(<u>**id**</u>, username, email, full_name, password_hash, available_balance, invested_balance, created_at, updated_at)
    8   - Candidate keys: `{id}`, `{username}`, `{email}`. `UNIQUE(username)`, `UNIQUE(email)`.
     11Notation: **bold** = primary key, *italic* = foreign key. After each foreign key
     12comes the ER relationship it implements.
     13
     14- **Users**(<u>**id**</u>, username, email, full_name, password_hash, available_balance, invested_balance, reserved_balance, created_at, updated_at)
     15  - Entity set `Users`. Candidate keys: `{id}`, `{username}`, `{email}`. `UNIQUE(username)`, `UNIQUE(email)`.
    916- **Crypto**(<u>**id**</u>, symbol, name, created_at)
    10   - Candidate keys: `{id}`, `{symbol}`. `UNIQUE(symbol)`.
    11 - **Markets**(<u>**id**</u>, *crypto_id*, quote_currency, is_active, created_at)
    12   - Candidate keys: `{id}`, `{crypto_id, quote_currency}`. `UNIQUE(crypto_id, quote_currency)`.
    13 - **Holdings**(<u>**id**</u>, *user_id*, *crypto_id*, quantity, reserved_quantity, avg_price, created_at, updated_at)
    14   - Transformation of the M:N relationship `Holds`. Candidate keys: `{id}` and
    15     `{user_id, crypto_id}` — the latter is the relationship's own key and is
    16     enforced with `UNIQUE(user_id, crypto_id)`. `id` was chosen as PK for
    17     consistency with the other relations.
     17  - Entity set `Cryptos`. Candidate keys: `{id}`, `{symbol}`. `UNIQUE(symbol)`.
     18- **Markets**(<u>**id**</u>, *crypto_id* [`QuotedOn`], quote_currency, is_active, created_at)
     19  - Entity set `Markets`. Candidate keys: `{id}`, `{crypto_id, quote_currency}`
     20    (the model's rule "a crypto is quoted at most once per currency"),
     21    enforced with `UNIQUE(crypto_id, quote_currency)`.
     22- **Holdings**(<u>**id**</u>, *user_id* [`Holds`], *crypto_id* [`PositionIn`], quantity, reserved_quantity, avg_price, created_at, updated_at)
     23  - Entity set `Holdings`. Candidate keys: `{id}` and `{user_id, crypto_id}`
     24    (the model's rule "one holding per user and crypto"), enforced with
     25    `UNIQUE(user_id, crypto_id)`.
    1826  - `avg_price` is `NOT NULL DEFAULT 0 CHECK (avg_price >= 0)`.
    1927  - `reserved_quantity` is `NOT NULL DEFAULT 0 CHECK (reserved_quantity >= 0
    … …  
    2331    needed, in `v_portfolio` as `available_quantity` and in the sell path of
    2432    [UseCase0005](../P3-UseCaseModel/UseCase0005.md). See
    25     [ERModel](../P1-ConceptualModel/ERModel.md#holds--users-m--cryptos-n-partial-on-both-sides-with-attributes)
     33    [ERModel](../P1-ConceptualModel/ERModel.md#holdings)
    2634    for why this mirrors `available_balance`/`invested_balance` on `Users`.
    27 - **Orders**(<u>**id**</u>, *user_id*, *market_id*, side, type, status, quantity, price, placed_at, executed_at)
    28   - `side ∈ {buy, sell}`, `type ∈ {market, limit}`, `status ∈ {open, executed, cancelled}`.
    29 - **Transactions**(<u>**id**</u>, *user_id*, type, amount, currency, *related_order*, created_at, description)
    30   - `type ∈ {deposit, buy, sell, fee}`.
    31 - **MarketTrades**(<u>**id**</u>, *market_id*, executed_at, price, quantity, side, source)
    32 - **MarketCandles**(<u>**id**</u>, *market_id*, timeframe, open, high, low, close, volume, candle_time)
    33   - `UNIQUE(market_id, timeframe, candle_time)`.
    34 - **Watchlists**(<u>**id**</u>, *user_id*, name, created_at)
    35 - **WatchlistItems**(<u>**id**</u>, *watchlist_id*, *crypto_id*, added_at)
    36   - Transformation of the M:N relationship `Contains`. Candidate keys: `{id}`
    37     and `{watchlist_id, crypto_id}`, the latter enforced with
    38     `UNIQUE(watchlist_id, crypto_id)`.
     35- **Orders**(<u>**id**</u>, *user_id* [`Places`], *market_id* [`PlacedOn`], side, type, status, quantity, filled_quantity, price, placed_at, executed_at)
     36  - Entity set `Orders`. `side ∈ {buy, sell}`, `type ∈ {market, limit}`,
     37    `status ∈ {open, partially_filled, executed, cancelled}`,
     38    `0 ≤ filled_quantity ≤ quantity`.
     39- **Transactions**(<u>**id**</u>, *user_id* [`Records`], type, amount, currency, *related_order* [`Settles`], created_at, description)
     40  - Entity set `Transactions`. `type ∈ {deposit, buy, sell, fee}`.
     41    `related_order` is nullable (see below).
     42- **MarketTrades**(<u>**id**</u>, *market_id* [`Fills`], executed_at, price, quantity, side, source, *buy_order_id* [`FillsBuy`], *sell_order_id* [`FillsSell`])
     43  - Entity set `MarketTrades`. `buy_order_id` and `sell_order_id` are both
     44    nullable (see below).
     45- **OrderEvents**(<u>**id**</u>, *order_id* [`Logs`], event_type, quantity, price, status_after, created_at)
     46  - Entity set `OrderEvents`. `event_type ∈ {placed, partially_filled, filled, cancelled}`.
     47- **MarketCandles**(<u>**id**</u>, *market_id* [`Aggregates`], timeframe, open, high, low, close, volume, candle_time)
     48  - Entity set `MarketCandles`. Candidate keys: `{id}`, `{market_id, timeframe,
     49    candle_time}` (the model's rule "one candle per market, timeframe and
     50    bucket"), enforced with `UNIQUE(market_id, timeframe, candle_time)`.
     51- **Watchlists**(<u>**id**</u>, *user_id* [`Owns`], name, created_at)
     52  - Entity set `Watchlists`.
     53- **WatchlistItems**(<u>**id**</u>, *watchlist_id* [`Contains`], *crypto_id* [`Lists`], added_at)
     54  - Entity set `WatchlistItems`. Candidate keys: `{id}` and `{watchlist_id,
     55    crypto_id}` (the model's rule "an asset at most once per list"), enforced
     56    with `UNIQUE(watchlist_id, crypto_id)`.
    3957
    4058### Transformation method used
    4159
    42 **Partial transformation.** Applied as follows:
    43 
    44 - Each of the 8 entity sets in [ERModel](../P1-ConceptualModel/ERModel.md) becomes one table, keeping
    45   its UUID (or serial) primary key.
    46 - Each **1:N relationship without attributes** is transformed by adding the
    47   parent's primary key as a foreign-key column on the child table — the "N"
    48   side. This is where every foreign key in the schema comes from, and it is why
    49   no foreign keys appear in the ER diagram itself:
    50   `QuotedOn` → `markets.crypto_id`, `PlacedOn` → `orders.market_id`,
    51   `Places` → `orders.user_id`, `Records` → `transactions.user_id`,
    52   `Settles` → `transactions.related_order`, `Fills` → `market_trades.market_id`,
    53   `Aggregates` → `market_candles.market_id`, `Owns` → `watchlists.user_id`.
    54 - Each **M:N relationship** becomes its own table holding the two foreign keys
    55   plus the relationship's own attributes: `Holds` → `holdings`,
    56   `Contains` → `watchlist_items`. The pair of foreign keys is the relationship's
    57   key and is enforced as a `UNIQUE` constraint in both tables.
    58 - **Total participation** in the ER model becomes `NOT NULL` on the
    59   corresponding foreign key; partial participation stays nullable. `Settles` is
    60   partial on both sides, which is exactly why `transactions.related_order` is
    61   the one nullable foreign key in the schema — a deposit has no originating
    62   order.
     60**Partial transformation.** The model has 11 entity sets and 15 relationships.
     61Every relationship is binary and 1:N with no attributes of its own (the two M:N
     62relationships of earlier versions, `Holds` and `Contains`, were corrected into
     63the entity sets `Holdings` and `WatchlistItems` in v05). The rules:
     64
     65- **Each entity set becomes one relation**, with its own attributes and its
     66  own key `id` as primary key. 11 entity sets → 11 relations.
     67- **Each 1:N relationship becomes one foreign key** on the relation of the "N"
     68  side, pointing to the primary key of the "1" side. No relationship gets its
     69  own table, because none is M:N and none has attributes. 15 relationships →
     70  15 foreign keys:
     71
     72  | ER relationship | 1 side → N side | Foreign key | Participation of the N side | `NULL`? |
     73  |---|---|---|---|---|
     74  | `QuotedOn`   | Cryptos → Markets             | `markets.crypto_id`            | total   | `NOT NULL` |
     75  | `PlacedOn`   | Markets → Orders              | `orders.market_id`             | total   | `NOT NULL` |
     76  | `Places`     | Users → Orders                | `orders.user_id`               | total   | `NOT NULL` |
     77  | `Records`    | Users → Transactions          | `transactions.user_id`         | total   | `NOT NULL` |
     78  | `Settles`    | Orders → Transactions         | `transactions.related_order`   | partial | nullable |
     79  | `Fills`      | Markets → MarketTrades        | `market_trades.market_id`      | total   | `NOT NULL` |
     80  | `FillsBuy`   | Orders → MarketTrades         | `market_trades.buy_order_id`   | partial | nullable |
     81  | `FillsSell`  | Orders → MarketTrades         | `market_trades.sell_order_id`  | partial | nullable |
     82  | `Logs`       | Orders → OrderEvents          | `order_events.order_id`        | total   | `NOT NULL` |
     83  | `Aggregates` | Markets → MarketCandles       | `market_candles.market_id`     | total   | `NOT NULL` |
     84  | `Owns`       | Users → Watchlists            | `watchlists.user_id`           | total   | `NOT NULL` |
     85  | `Holds`      | Users → Holdings              | `holdings.user_id`             | total   | `NOT NULL` |
     86  | `PositionIn` | Cryptos → Holdings            | `holdings.crypto_id`           | total   | `NOT NULL` |
     87  | `Contains`   | Watchlists → WatchlistItems   | `watchlist_items.watchlist_id` | total   | `NOT NULL` |
     88  | `Lists`      | Cryptos → WatchlistItems      | `watchlist_items.crypto_id`    | total   | `NOT NULL` |
     89
     90- **Participation decides `NULL`.** Total participation of the N side means
     91  every row must reference a parent, so the foreign key is `NOT NULL`. Partial
     92  participation leaves it nullable. There are exactly three partial ones:
     93  `Settles` (a deposit has no originating order), and `FillsBuy` / `FillsSell`
     94  (a trade against the simulated market has no user order on that side).
     95  Partial participation of the **1** side (for example, a user with no orders)
     96  needs no column at all. It simply means no row points at that parent.
     97- **Uniqueness rules of the model become `UNIQUE` constraints.** The four
     98  rules the model states in words ("a crypto quoted once per currency", "one
     99  candle per market, timeframe and bucket", "one holding per user and crypto",
     100  "an asset once per list") involve a relationship, so Chen notation cannot
     101  draw them as keys. After transformation, the relationship is a foreign-key
     102  column, and each rule becomes an ordinary composite `UNIQUE` constraint, i.e. a
     103  second candidate key.
     104
     105Nothing in the schema comes from anywhere else. Every column is either an ER
     106attribute or the foreign key of one listed relationship.
    63107
    64108### Normalisation
    65109
    66 > **Validated in P5.** [Normalization](../P5-Normalization/Normalization.md) derives this
    67 > exact schema independently — starting only from a single de-normalized relation of every
    68 > model attribute and its functional dependencies, with no reference to the ER-to-relational
    69 > transformation below — and shows it decomposes to **BCNF**, one normal form stronger than
    70 > the 3NF claimed here. The two designs agree relation for relation and key for key, so
    71 > nothing here changed as a result; see that page's
    72 > [discussion](../P5-Normalization/Normalization.md#discussion) for what the one real
    73 > difference is (`avg_price`, a stored derived value, not a normalisation issue) and why this
    74 > design is still the one used from P5 onward.
    75 
    76 All relations are in **3NF**:
     110> **Checked in P5.** [Normalization](../P5-Normalization/Normalization.md)
     111> starts from a single de-normalized relation containing only the attributes
     112> of the ER model and the functional dependencies that follow from its rules.
     113> It decomposes that relation step by step to BCNF and arrives at these same 11
     114> relations, with one deliberate difference: `transactions.user_id` (see the last
     115> bullet below). The comparison is in the *Discussion* section at the end of that
     116> page.
     117
     118All relations except `transactions` are in **BCNF**, as P5 shows. `transactions`
     119is in 2NF but not in 3NF, because of the deliberately kept `user_id` (last
     120bullet):
    77121
    78122- Every attribute is atomic (no repeating groups, no composite fields).
    79 - No partial dependency exists because every primary key is a single UUID column.
    80 - No transitive dependency exists: every non-key attribute depends directly on the row identifier. For example, `holdings.quantity` depends on `holdings.id`, not on `user_id` via some intermediate.
     123- No partial dependency exists: every candidate key is either the single
     124  column `id` or a composite key (`{user_id, crypto_id}`, …) on which no
     125  non-key attribute depends only partially.
     126- No transitive dependency exists, except `transactions.user_id` (last bullet):
     127  every other non-key attribute depends directly on the row's own entity, never
     128  on another entity reached through a foreign key.
     129  For example, `holdings.quantity` depends on `holdings.id`, and nothing about
     130  the user or the crypto is copied into `holdings`.
    81131- `avg_price` in `Holdings` is a **derived value** cached for performance (it is
    82132  the weighted-average entry price across all `buy` transactions for that
    … …  
    95145  ("available") is the derived value here, and it is never stored, only
    96146  computed where it is needed.
     147- `transactions.user_id` is kept **deliberately**, although for an entry that
     148  settles an order it repeats that order's user (`related_order → user_id`, a
     149  transitive dependency). A deposit has no order (`Settles` is partial), so
     150  `user_id` is the only way to record whose deposit it is. For entries with an
     151  order, the only code that sets `related_order` (the buy and sell inserts in
     152  `advanced_db.sql`) writes both from the same order row. No database
     153  constraint enforces this.
    97154
    98155### Reservation and the order lifecycle
    … …  
    109166`SELECT … FOR UPDATE` locking that already protected `users.available_balance`
    110167on the buy path is what makes two concurrent sell orders against the same
    111 holding serialize correctly instead of racing.
     168holding serialize correctly instead of racing. The cash side of a buy order
     169(`users.reserved_balance`), `orders.filled_quantity` and `order_events` were
     170added in P7; see
     171[AdvancedDatabaseDevelopment](../P7-AdvancedDatabaseDevelopment/AdvancedDatabaseDevelopment.md).
    112172
    113173## DDL script
    114174
    115 The script that creates the entire schema is [`../server/db/schema_creation.sql`](../../server/db/schema_creation.sql). It is idempotent: it drops and recreates the `project` schema every run, so it works on an empty database and on a database that already has the schema.
     175The script that creates the schema is [`../server/db/schema_creation.sql`](../../server/db/schema_creation.sql). It is idempotent: it drops and recreates the `project` schema every run, so it works on an empty database and on a database that already has the schema.
    116176
    117177The script creates:
    118 - 10 tables with check constraints, primary keys, foreign keys and unique constraints.
    119 - 5 performance indexes.
     178- 10 of the 11 tables, with check constraints, primary keys, foreign keys and unique constraints.
     179- 8 performance indexes.
    120180- 2 views: `v_latest_prices` (latest trade price per market) and `v_portfolio` (per-user holdings valuation with unrealised P/L, plus `reserved_quantity` and the derived `available_quantity`).
     181
     182The 11th table, `order_events`, is created by
     183[`../server/db/advanced_db.sql`](../../server/db/advanced_db.sql) together with
     184the P7 triggers that fill it. `./eduberza -init` runs both scripts in that
     185order, so a freshly initialised database always has all 11 tables and all 15
     186foreign keys.
    121187
    122188## DML script (sample data)
    … …  
    132198## Relational diagram
    133199
    134 ![relational_schema](relational_schema.jpg)
    135 
    136 Generated in **Pgadmin** from the **live** `project` schema, in crow's-foot
    137 notation — not drawn by hand, so it is evidence that the deployed database
    138 actually matches the design described above. Each box is a table with its
    139 columns and declared types; key icons mark primary keys and the arrowed lines
    140 are the 12 declared foreign keys.
     200![relational_diagram_v4](relational_diagram_v4.png)
     201
     202Generated in **DBeaver** from the **live** `project` schema (after
     203`./eduberza -init`), not drawn by hand, so it shows what the deployed database
     204actually contains. Each box is a table with its columns; the key icon marks the
     205primary key, and the lines are the 15 declared foreign keys. The two foreign keys from
     206`market_trades` to `orders` (`buy_order_id`, `sell_order_id`) connect the same
     207two boxes, so DBeaver draws them on top of each other as one line.
     208
     209The tables are arranged in the **same positions** as the entity sets in
     210`ERModel_v05.png`, so the two can be compared directly:
     211
     212- every rectangle of the ER diagram is one table in the same place;
     213- every diamond of the ER diagram is one foreign-key line between the same two
     214  boxes. The dot is on the referencing ("N") table, next to the foreign-key
     215  column;
     216- a double (total) line in the ER diagram is a `NOT NULL` foreign key, drawn by
     217  DBeaver as a solid line. The three single lines on the N side (`Settles`,
     218  `FillsBuy`, `FillsSell`) are the three nullable foreign keys, which DBeaver
     219  draws dashed, with a hollow diamond on the `orders` side. The table under
     220  [Transformation method used](#transformation-method-used) lists all 15.
     221
     222Earlier images (`relational_schema.jpg`, `relational_schema_v2.png`,
     223`relational_schema_v3.png`) were exported from pgAdmin, with a different layout
     224and from an older schema. They are kept only as history.
    141225
    142226### How to regenerate it
    143227
    144 **With pgAdmin 4**, if DBeaver is unavailable — it reads the live schema the same
    145 way, so the result is equivalent in substance:
    146 
    147 1. Connect to the project database.
    148 2. Right-click the database → **ERD For Database** (or open a blank ERD and drag
    149    the `project` tables in).
    150 3. Arrange the tables to mirror `ERModel_v03.png`.
    151 4. **Download image** → PNG, then convert:
    152    `convert relational_schema.png relational_schema.jpg`
    153 
     2281. Initialise the database: `./eduberza -init` (runs `schema_creation.sql` and
     229   `advanced_db.sql`, so `order_events` is included).
     2302. In DBeaver, connect to the project database and expand
     231   *Schemas → project → Tables*.
     2323. Select all 11 tables → right-click → **View Diagram** (or create a new ER
     233   diagram and drag the tables in).
     2344. Drag each table to the position of its entity set in `ERModel_v05.png`:
     235
     236   ```
     237            column 1          column 2        column 3        column 4
     238   row 1    watchlist_items   crypto          markets         market_candles
     239   row 2    watchlists        holdings        .               market_trades
     240   row 3    users             .               orders          .
     241   row 4    .                 transactions    order_events    .
     242   ```
     243
     244   Leave the empty cells (`.`) empty. They are where the relationship
     245   diamonds are in the ER diagram, so the foreign-key lines will run through
     246   the same gaps.
     247
     2485. Right-click the canvas → **Export diagram** → PNG, saved as
     249   `relational_diagram_v4.png` in this folder.
Note: See TracChangeset for help on using the changeset viewer.