Add migration to fix exchange events primary keys (#1593)

* Add migration to fix exchange events primary key

* correct comment: "foreign key" -> "primary key"

* Refine hack to handle only the expected error

* Add tx hash to erc20 approval events primary key

7 files changed, 79 insertions, 7 deletions

diff --git a/packages/pipeline/migrations/1549479172800-AddTxHashToExchangeEventPrimaryKey.ts b/packages/pipeline/migrations/1549479172800-AddTxHashToExchangeEventPrimaryKey.ts
new file mode 100644
index 000000000..d6ea6c47b
--- /dev/null
+++ b/packages/pipeline/migrations/1549479172800-AddTxHashToExchangeEventPrimaryKey.ts
@@ -0,0 +1,35 @@
+import { MigrationInterface, QueryRunner } from 'typeorm';
+
+const tableNames = ['exchange_cancel_events', 'exchange_cancel_up_to_events', 'exchange_fill_events'];
+
+const oldPrimaryColumns = ['contract_address', 'log_index', 'block_number'];
+
+const newPrimaryColumns = ['transaction_hash'];
+
+async function updatePrimaryKeysAsync(queryRunner: QueryRunner, columnNames: string[]): Promise<void> {
+    for (const tableName of tableNames) {
+        const table = await queryRunner.getTable(`raw.${tableName}`);
+        if (table === undefined) {
+            throw new Error(`Couldn't get table 'raw.${tableName}'`);
+        }
+        const columns = [];
+        for (const columnName of columnNames) {
+            const column = table.findColumnByName(columnName);
+            if (column === undefined) {
+                throw new Error(`Couldn't get column '${columnName}' from table 'raw.${tableName}'`);
+            }
+            columns.push(column);
+        }
+        await queryRunner.updatePrimaryKeys(table, columns);
+    }
+}
+
+export class AddTxHashToExchangeEventPrimaryKey1549479172800 implements MigrationInterface {
+    public async up(queryRunner: QueryRunner): Promise<any> {
+        await updatePrimaryKeysAsync(queryRunner, oldPrimaryColumns.concat(newPrimaryColumns));
+    }
+
+    public async down(queryRunner: QueryRunner): Promise<any> {
+        await updatePrimaryKeysAsync(queryRunner, oldPrimaryColumns);
+    }
+}
diff --git a/packages/pipeline/migrations/1549499426238-AddTxHashToERC20ApprovalEventPrimaryKey.ts b/packages/pipeline/migrations/1549499426238-AddTxHashToERC20ApprovalEventPrimaryKey.ts
new file mode 100644
index 000000000..874713e67
--- /dev/null
+++ b/packages/pipeline/migrations/1549499426238-AddTxHashToERC20ApprovalEventPrimaryKey.ts
@@ -0,0 +1,31 @@
+import { MigrationInterface, QueryRunner } from 'typeorm';
+
+const oldPrimaryColumns = ['token_address', 'log_index', 'block_number'];
+
+const newPrimaryColumns = ['transaction_hash'];
+
+async function updatePrimaryKeysAsync(queryRunner: QueryRunner, columnNames: string[]): Promise<void> {
+    const table = await queryRunner.getTable(`raw.erc20_approval_events`);
+    if (table === undefined) {
+        throw new Error(`Couldn't get table 'raw.erc20_approval_events'`);
+    }
+    const columns = [];
+    for (const columnName of columnNames) {
+        const column = table.findColumnByName(columnName);
+        if (column === undefined) {
+            throw new Error(`Couldn't get column '${columnName}' from table 'raw.erc20_approval_events'`);
+        }
+        columns.push(column);
+    }
+    await queryRunner.updatePrimaryKeys(table, columns);
+}
+
+export class AddTxHashToERC20ApprovalEventPrimaryKey1549499426238 implements MigrationInterface {
+    public async up(queryRunner: QueryRunner): Promise<any> {
+        await updatePrimaryKeysAsync(queryRunner, oldPrimaryColumns.concat(newPrimaryColumns));
+    }
+
+    public async down(queryRunner: QueryRunner): Promise<any> {
+        await updatePrimaryKeysAsync(queryRunner, oldPrimaryColumns);
+    }
+}
diff --git a/packages/pipeline/src/entities/erc20_approval_event.ts b/packages/pipeline/src/entities/erc20_approval_event.ts
index 69cdfcb0b..ee5e621d2 100644
--- a/packages/pipeline/src/entities/erc20_approval_event.ts
+++ b/packages/pipeline/src/entities/erc20_approval_event.ts
@@ -15,7 +15,7 @@ export class ERC20ApprovalEvent {
     @Column({ name: 'raw_data' })
     public rawData!: string;
 
-    @Column({ name: 'transaction_hash' })
+    @PrimaryColumn({ name: 'transaction_hash' })
     public transactionHash!: string;
     @Column({ name: 'owner_address' })
     public ownerAddress!: string;
diff --git a/packages/pipeline/src/entities/exchange_cancel_event.ts b/packages/pipeline/src/entities/exchange_cancel_event.ts
index 38f99c903..a86194920 100644
--- a/packages/pipeline/src/entities/exchange_cancel_event.ts
+++ b/packages/pipeline/src/entities/exchange_cancel_event.ts
@@ -15,7 +15,7 @@ export class ExchangeCancelEvent {
     @Column({ name: 'raw_data' })
     public rawData!: string;
 
-    @Column({ name: 'transaction_hash' })
+    @PrimaryColumn({ name: 'transaction_hash' })
     public transactionHash!: string;
     @Column({ name: 'maker_address' })
     public makerAddress!: string;
diff --git a/packages/pipeline/src/entities/exchange_cancel_up_to_event.ts b/packages/pipeline/src/entities/exchange_cancel_up_to_event.ts
index 27580305e..f24aea23a 100644
--- a/packages/pipeline/src/entities/exchange_cancel_up_to_event.ts
+++ b/packages/pipeline/src/entities/exchange_cancel_up_to_event.ts
@@ -15,7 +15,7 @@ export class ExchangeCancelUpToEvent {
     @Column({ name: 'raw_data' })
     public rawData!: string;
 
-    @Column({ name: 'transaction_hash' })
+    @PrimaryColumn({ name: 'transaction_hash' })
     public transactionHash!: string;
     @Column({ name: 'maker_address' })
     public makerAddress!: string;
diff --git a/packages/pipeline/src/entities/exchange_fill_event.ts b/packages/pipeline/src/entities/exchange_fill_event.ts
index 9b7727615..52111711e 100644
--- a/packages/pipeline/src/entities/exchange_fill_event.ts
+++ b/packages/pipeline/src/entities/exchange_fill_event.ts
@@ -16,7 +16,7 @@ export class ExchangeFillEvent {
     @Column({ name: 'raw_data' })
     public rawData!: string;
 
-    @Column({ name: 'transaction_hash' })
+    @PrimaryColumn({ name: 'transaction_hash' })
     public transactionHash!: string;
     @Column({ name: 'maker_address' })
     public makerAddress!: string;
diff --git a/packages/pipeline/src/scripts/pull_exchange_events.ts b/packages/pipeline/src/scripts/pull_exchange_events.ts
index f8ce4038d..c2c56da6b 100644
--- a/packages/pipeline/src/scripts/pull_exchange_events.ts
+++ b/packages/pipeline/src/scripts/pull_exchange_events.ts
@@ -112,15 +112,20 @@ async function saveIndividuallyWithFallbackAsync<T extends ExchangeEvent>(
     events: T[],
 ): Promise<void> {
     // Note(albrow): This is a temporary hack because `save` is not working as
-    // documented and is causing a foreign key constraint violation. Hopefully
+    // documented and is causing a primary key constraint violation. Hopefully
     // can remove later because this "poor man's upsert" implementation operates
     // on one event at a time and is therefore much slower.
     for (const event of events) {
         try {
             // First try an insert.
             await repository.insert(event);
-        } catch {
-            // If it fails, assume it was a foreign key constraint error and try
+        } catch (err) {
+            if (err.message.includes('duplicate key value violates unique constraint')) {
+                logUtils.log("Ignore the preceeding INSERT failure; it's not unexpected");
+            } else {
+                throw err;
+            }
+            // If it fails, assume it was a primary key constraint error and try
             // doing an update instead.
             // Note(albrow): Unfortunately the `as any` hack here seems
             // required. I can't figure out how to convince the type-checker
@@ -132,6 +137,7 @@ async function saveIndividuallyWithFallbackAsync<T extends ExchangeEvent>(
                     contractAddress: event.contractAddress,
                     blockNumber: event.blockNumber,
                     logIndex: event.logIndex,
+                    transactionHash: event.transactionHash,
                 } as any,
                 event as any,
             );