Merge pull request #507 from FlipsideCrypto/DAT2-141-docs

DAT2-141/upd intents table doc

models/descriptions/tables/defi__ez_intents.md

@@ -1,16 +1,100 @@
 {% docs defi__ez_intents %}
 
 ## Description
-This table provides an enhanced view of all intent-based transactions on the NEAR Protocol blockchain, combining raw intent data with token metadata, pricing information, and cross-chain mapping details. The table includes both NEP-245 and DIP-4 standard intents with decimal adjustments, USD values, and comprehensive token context. This easy view eliminates the need for complex joins and calculations, providing ready-to-use data for intent-based trading analysis and MEV protection mechanism evaluation.
+This table provides an enhanced view of all intent-based transactions on the NEAR Protocol blockchain, combining raw intent data with token metadata, pricing information, and cross-chain mapping details. The table includes both NEP-245 and DIP-4 standard intents with decimal adjustments, USD values, and comprehensive token context.
+
+**CRITICAL**: Each intent consists of multiple transaction steps (rows) representing the complete transfer flow: input (user → solver), output (solver → user), fees, and optional routing. Each unique intent is identified by `tx_hash` + `log_index`. Summing all `amount_usd` values will overcount volume by 2-3x because it counts input + output + fees separately.
+
+## How to Use This Table
+
+### Volume Calculation (Recommended Method)
+**Use MAX per intent** to avoid 2-3x overcounting:
+
+```sql
+WITH intent_volumes AS (
+    SELECT
+        tx_hash,
+        log_index,
+        MAX(amount_usd) as intent_volume
+    FROM near.defi.ez_intents
+    WHERE amount_usd > 0.01  -- Exclude dust/fees
+    GROUP BY tx_hash, log_index  -- BOTH fields required
+)
+SELECT SUM(intent_volume) as total_volume
+FROM intent_volumes;
+```
+
+**Why MAX works**: Input ≈ output amounts (differ only by small fees). MAX captures the true swap size while ignoring dust.
+
+### Common Query Patterns
+
+**Daily volume by token:**
+```sql
+WITH intent_volumes AS (
+    SELECT
+        DATE_TRUNC('day', block_timestamp) as date,
+        tx_hash, log_index, symbol,
+        MAX(amount_usd) as volume
+    FROM near.defi.ez_intents
+    WHERE amount_usd > 0.01
+    GROUP BY date, tx_hash, log_index, symbol
+)
+SELECT date, symbol, SUM(volume) as daily_volume
+FROM intent_volumes
+GROUP BY date, symbol;
+```
+
+**Cross-chain flows:**
+```sql
+WITH flows AS (
+    SELECT
+        tx_hash, log_index,
+        MIN(blockchain) as source_chain,
+        MAX(blockchain) as dest_chain,
+        MAX(amount_usd) as volume
+    FROM near.defi.ez_intents
+    WHERE amount_usd > 0.01
+    GROUP BY tx_hash, log_index
+)
+SELECT source_chain, dest_chain, SUM(volume) as total_volume
+FROM flows
+GROUP BY source_chain, dest_chain;
+```
+
+**Top users:**
+```sql
+WITH users AS (
+    SELECT
+        old_owner_id as user,
+        tx_hash, log_index,
+        MAX(amount_usd) as volume
+    FROM near.defi.ez_intents
+    WHERE amount_usd > 0.01
+    GROUP BY user, tx_hash, log_index
+)
+SELECT user, SUM(volume) as total_volume
+FROM users
+GROUP BY user;
+```
+
+### Validation
+Check if you're calculating correctly:
+```sql
+SELECT
+    AVG(SUM(amount_usd) / NULLIF(MAX(amount_usd), 0)) as avg_ratio
+FROM near.defi.ez_intents
+WHERE amount_usd > 0.01
+GROUP BY tx_hash, log_index;
+```
+**Expected**: ~1.5-2.0. If higher, you're likely overcounting.
 
 ## Key Use Cases
-- Intent-based trading analysis with accurate decimal-adjusted amounts and USD values
+- Intent-based trading volume analysis with accurate calculations
-- MEV protection mechanism effectiveness analysis with economic impact
+- Cross-chain swap tracking and flow analysis
-- User behavior analysis in intent-driven protocols with value context
+- User behavior analysis in intent-driven protocols
-- Cross-protocol intent execution tracking and performance comparison
+- Referral program effectiveness measurement
-- Intent fulfillment rate analysis and optimization with value calculations
+- Intent fulfillment rate analysis (use `receipt_succeeded`)
-- Referral program analysis and user acquisition tracking
+- MEV protection mechanism evaluation
-- Intent-based DeFi protocol performance monitoring and benchmarking
 
 ## Important Relationships
 - Enhances `defi.fact_intents` with pricing and metadata from `price.ez_prices_hourly`
@@ -20,12 +104,15 @@ This table provides an enhanced view of all intent-based transactions on the NEA
 - Powers cross-protocol analysis with other DeFi activities
 
 ## Commonly-used Fields
-- `amount` and `amount_usd`: Essential for accurate intent value analysis
+- `tx_hash` + `log_index`: **Unique intent identifier** - MUST group by both for accurate volume calculations
-- `block_timestamp`: Primary field for time-series analysis and trend detection
+- `amount_usd`: Intent value in USD - use with MAX() per intent to avoid overcounting
-- `log_event`: Critical for intent type classification and analysis
+- `symbol`: Token symbol for token-specific analysis (USDT, NEAR, ETH, etc.)
-- `owner_id`: Important for user behavior analysis and intent tracking
+- `blockchain`: Source/destination chain (near, eth, bsc, sol, etc.) for cross-chain flow analysis
-- `token_id` and `symbol`: Essential for token-specific intent analysis
+- `old_owner_id` / `new_owner_id`: Sender/recipient addresses - use `old_owner_id` to identify intent initiator
-- `referral`: Important for referral program analysis and user acquisition
+- `log_event_index`: Step order within intent - useful for understanding transaction flow
-- `receipt_succeeded`: Critical for intent fulfillment rate analysis
+- `block_timestamp`: Time of transaction - use for time-series analysis and filtering
+- `receipt_succeeded`: Transaction success status - filter to TRUE for successful intents only
+- `referral`: Referral address for referral program analysis
+- `fee_amount_usd`: Fee collected in USD - separate from main intent volume
 
 {% enddocs %}