Search in sources :

Example 1 with DependableFinder

use of org.apache.derby.catalog.DependableFinder in project derby by apache.

the class DDLConstantAction method addColumnDependencies.

/**
 * Add dependencies of a column on providers. These can arise if a generated column depends
 * on a user created function.
 */
protected void addColumnDependencies(LanguageConnectionContext lcc, DataDictionary dd, TableDescriptor td, ColumnInfo ci) throws StandardException {
    ProviderInfo[] providers = ci.providers;
    if (providers != null) {
        DependencyManager dm = dd.getDependencyManager();
        ContextManager cm = lcc.getContextManager();
        int providerCount = providers.length;
        ColumnDescriptor cd = td.getColumnDescriptor(ci.name);
        DefaultDescriptor defDesc = cd.getDefaultDescriptor(dd);
        for (int px = 0; px < providerCount; px++) {
            ProviderInfo pi = providers[px];
            DependableFinder finder = pi.getDependableFinder();
            UUID providerID = pi.getObjectId();
            Provider provider = (Provider) finder.getDependable(dd, providerID);
            dm.addDependency(defDesc, provider, cm);
        }
    // end loop through providers
    }
}
Also used : DependableFinder(org.apache.derby.catalog.DependableFinder) ProviderInfo(org.apache.derby.iapi.sql.depend.ProviderInfo) ContextManager(org.apache.derby.iapi.services.context.ContextManager) ColumnDescriptor(org.apache.derby.iapi.sql.dictionary.ColumnDescriptor) DependencyManager(org.apache.derby.iapi.sql.depend.DependencyManager) DefaultDescriptor(org.apache.derby.iapi.sql.dictionary.DefaultDescriptor) UUID(org.apache.derby.catalog.UUID) Provider(org.apache.derby.iapi.sql.depend.Provider)

Example 2 with DependableFinder

use of org.apache.derby.catalog.DependableFinder in project derby by apache.

the class dblook_test method getDependsData.

/* **********************************************
	 * getDependsData:
	 * Forms a string containing detailed information
	 * about a row in the SYSDEPENDS table, and returns
	 * that string.
	 * @param rs Result set with SYSDEPENDS rows; current
	 *  row is the one for which we're getting the
	 *  data.
	 * @param conn Connection to the database being
	 *   examined.
	 * @param idToNameMap mapping of object ids to names
	 *  for the database in question.
	 * @return Schema, type and name of both the Provider
	 *   and the Dependent for the current row of
	 *   SYSDEPENDS have been returned as a string.
	 ****/
private String getDependsData(ResultSet rs, Connection conn, HashMap idToNameMap) throws Exception {
    DependableFinder dep = (DependableFinder) rs.getObject(3);
    DependableFinder prov = (DependableFinder) rs.getObject(5);
    String depType = dep.getSQLObjectType();
    String provType = prov.getSQLObjectType();
    Statement dependsStmt = conn.createStatement();
    StringBuffer dependsData = new StringBuffer();
    dependsData.append(getHiddenDependsData(depType, rs.getString(2), dependsStmt, idToNameMap));
    dependsData.append(" -> ");
    dependsData.append(getHiddenDependsData(provType, rs.getString(4), dependsStmt, idToNameMap));
    return dependsData.toString();
}
Also used : DependableFinder(org.apache.derby.catalog.DependableFinder) Statement(java.sql.Statement)

Example 3 with DependableFinder

use of org.apache.derby.catalog.DependableFinder in project derby by apache.

the class AlterTableConstantAction method dropColumnFromTable.

/**
 * Workhorse for dropping a column from a table.
 *
 * This routine drops a column from a table, taking care
 * to properly handle the various related schema objects.
 *
 * The syntax which gets you here is:
 *
 *   ALTER TABLE tbl DROP [COLUMN] col [CASCADE|RESTRICT]
 *
 * The keyword COLUMN is optional, and if you don't
 * specify CASCADE or RESTRICT, the default is CASCADE
 * (the default is chosen in the parser, not here).
 *
 * If you specify RESTRICT, then the column drop should be
 * rejected if it would cause a dependent schema object
 * to become invalid.
 *
 * If you specify CASCADE, then the column drop should
 * additionally drop other schema objects which have
 * become invalid.
 *
 * You may not drop the last (only) column in a table.
 *
 * Schema objects of interest include:
 *  - views
 *  - triggers
 *  - constraints
 *    - check constraints
 *    - primary key constraints
 *    - foreign key constraints
 *    - unique key constraints
 *    - not null constraints
 *  - privileges
 *  - indexes
 *  - default values
 *
 * Dropping a column may also change the column position
 * numbers of other columns in the table, which may require
 * fixup of schema objects (such as triggers and column
 * privileges) which refer to columns by column position number.
 *
 * Indexes are a bit interesting. The official SQL spec
 * doesn't talk about indexes; they are considered to be
 * an imlementation-specific performance optimization.
 * The current Derby behavior is that:
 *  - CASCADE/RESTRICT doesn't matter for indexes
 *  - when a column is dropped, it is removed from any indexes
 *    which contain it.
 *  - if that column was the only column in the index, the
 *    entire index is dropped.
 *
 * @param   columnName the name of the column specfication in the ALTER
 *						statement-- currently we allow only one.
 * @exception StandardException 	thrown on failure.
 */
private void dropColumnFromTable(String columnName) throws StandardException {
    boolean cascade = (behavior == StatementType.DROP_CASCADE);
    // drop any generated columns which reference this column
    ColumnDescriptorList generatedColumnList = td.getGeneratedColumns();
    int generatedColumnCount = generatedColumnList.size();
    ArrayList<String> cascadedDroppedColumns = new ArrayList<String>();
    for (int i = 0; i < generatedColumnCount; i++) {
        ColumnDescriptor generatedColumn = generatedColumnList.elementAt(i);
        String[] referencedColumnNames = generatedColumn.getDefaultInfo().getReferencedColumnNames();
        int referencedColumnCount = referencedColumnNames.length;
        for (int j = 0; j < referencedColumnCount; j++) {
            if (columnName.equals(referencedColumnNames[j])) {
                String generatedColumnName = generatedColumn.getColumnName();
                // we're trying to drop
                if (!cascade) {
                    // 
                    throw StandardException.newException(SQLState.LANG_PROVIDER_HAS_DEPENDENT_OBJECT, dm.getActionString(DependencyManager.DROP_COLUMN), columnName, "GENERATED COLUMN", generatedColumnName);
                } else {
                    cascadedDroppedColumns.add(generatedColumnName);
                }
            }
        }
    }
    DataDescriptorGenerator ddg = dd.getDataDescriptorGenerator();
    int cascadedDrops = cascadedDroppedColumns.size();
    int sizeAfterCascadedDrops = td.getColumnDescriptorList().size() - cascadedDrops;
    // can NOT drop a column if it is the only one in the table
    if (sizeAfterCascadedDrops == 1) {
        throw StandardException.newException(SQLState.LANG_PROVIDER_HAS_DEPENDENT_OBJECT, dm.getActionString(DependencyManager.DROP_COLUMN), "THE *LAST* COLUMN " + columnName, "TABLE", td.getQualifiedName());
    }
    // now drop dependent generated columns
    for (int i = 0; i < cascadedDrops; i++) {
        String generatedColumnName = cascadedDroppedColumns.get(i);
        activation.addWarning(StandardException.newWarning(SQLState.LANG_GEN_COL_DROPPED, generatedColumnName, td.getName()));
        // 
        // We can only recurse 2 levels since a generation clause cannot
        // refer to other generated columns.
        // 
        dropColumnFromTable(generatedColumnName);
    }
    /*
         * Cascaded drops of dependent generated columns may require us to
         * rebuild the table descriptor.
         */
    td = dd.getTableDescriptor(tableId);
    ColumnDescriptor columnDescriptor = td.getColumnDescriptor(columnName);
    // We already verified this in bind, but do it again
    if (columnDescriptor == null) {
        throw StandardException.newException(SQLState.LANG_COLUMN_NOT_FOUND_IN_TABLE, columnName, td.getQualifiedName());
    }
    int size = td.getColumnDescriptorList().size();
    droppedColumnPosition = columnDescriptor.getPosition();
    FormatableBitSet toDrop = new FormatableBitSet(size + 1);
    toDrop.set(droppedColumnPosition);
    td.setReferencedColumnMap(toDrop);
    dm.invalidateFor(td, (cascade ? DependencyManager.DROP_COLUMN : DependencyManager.DROP_COLUMN_RESTRICT), lcc);
    // If column has a default we drop the default and any dependencies
    if (columnDescriptor.getDefaultInfo() != null) {
        dm.clearDependencies(lcc, columnDescriptor.getDefaultDescriptor(dd));
    }
    // then we need to drop the system-generated sequence backing it.
    if (columnDescriptor.isAutoincrement() && dd.checkVersion(DataDictionary.DD_VERSION_DERBY_10_11, null)) {
        DropTableConstantAction.dropIdentitySequence(dd, td, activation);
    }
    // columns which are used through REFERENCING clause
    for (TriggerDescriptor trd : dd.getTriggerDescriptors(td)) {
        // If we find that the trigger is dependent on the column being
        // dropped because column is part of trigger columns list, then
        // we will give a warning or drop the trigger based on whether
        // ALTER TABLE DROP COLUMN is RESTRICT or CASCADE. In such a
        // case, no need to check if the trigger action columns referenced
        // through REFERENCING clause also used the column being dropped.
        boolean triggerDroppedAlready = false;
        int[] referencedCols = trd.getReferencedCols();
        if (referencedCols != null) {
            int refColLen = referencedCols.length, j;
            boolean changed = false;
            for (j = 0; j < refColLen; j++) {
                if (referencedCols[j] > droppedColumnPosition) {
                    // Trigger is not defined on the column being dropped
                    // but the column position of trigger column is changing
                    // because the position of the column being dropped is
                    // before the the trigger column
                    changed = true;
                } else if (referencedCols[j] == droppedColumnPosition) {
                    // the trigger is defined on the column being dropped
                    if (cascade) {
                        trd.drop(lcc);
                        triggerDroppedAlready = true;
                        activation.addWarning(StandardException.newWarning(SQLState.LANG_TRIGGER_DROPPED, trd.getName(), td.getName()));
                    } else {
                        // otherwsie there would be unexpected behaviors
                        throw StandardException.newException(SQLState.LANG_PROVIDER_HAS_DEPENDENT_OBJECT, dm.getActionString(DependencyManager.DROP_COLUMN), columnName, "TRIGGER", trd.getName());
                    }
                    break;
                }
            }
            // drop column.
            if (j == refColLen && changed) {
                dd.dropTriggerDescriptor(trd, tc);
                for (j = 0; j < refColLen; j++) {
                    if (referencedCols[j] > droppedColumnPosition)
                        referencedCols[j]--;
                }
                trd.setReferencedCols(referencedCols);
                dd.addDescriptor(trd, sd, DataDictionary.SYSTRIGGERS_CATALOG_NUM, false, tc);
            }
        }
        // loop above, then move to next trigger
        if (triggerDroppedAlready)
            continue;
        // Column being dropped is not one of trigger columns. Check if
        // that column is getting used inside the trigger action through
        // REFERENCING clause. This can be tracked only for triggers
        // created in 10.7 and higher releases. Derby releases prior to
        // that did not keep track of trigger action columns used
        // through the REFERENCING clause.
        int[] referencedColsInTriggerAction = trd.getReferencedColsInTriggerAction();
        if (referencedColsInTriggerAction != null) {
            int refColInTriggerActionLen = referencedColsInTriggerAction.length, j;
            boolean changedColPositionInTriggerAction = false;
            for (j = 0; j < refColInTriggerActionLen; j++) {
                if (referencedColsInTriggerAction[j] > droppedColumnPosition) {
                    changedColPositionInTriggerAction = true;
                } else if (referencedColsInTriggerAction[j] == droppedColumnPosition) {
                    if (cascade) {
                        trd.drop(lcc);
                        triggerDroppedAlready = true;
                        activation.addWarning(StandardException.newWarning(SQLState.LANG_TRIGGER_DROPPED, trd.getName(), td.getName()));
                    } else {
                        // we'd better give an error if don't drop it,
                        throw StandardException.newException(SQLState.LANG_PROVIDER_HAS_DEPENDENT_OBJECT, dm.getActionString(DependencyManager.DROP_COLUMN), columnName, "TRIGGER", trd.getName());
                    }
                    break;
                }
            }
            // column has been actually dropped from the table descriptor.
            if (j == refColInTriggerActionLen && changedColPositionInTriggerAction) {
                dd.dropTriggerDescriptor(trd, tc);
                for (j = 0; j < refColInTriggerActionLen; j++) {
                    if (referencedColsInTriggerAction[j] > droppedColumnPosition)
                        referencedColsInTriggerAction[j]--;
                }
                trd.setReferencedColsInTriggerAction(referencedColsInTriggerAction);
                dd.addDescriptor(trd, sd, DataDictionary.SYSTRIGGERS_CATALOG_NUM, false, tc);
            }
        }
    }
    ConstraintDescriptorList csdl = dd.getConstraintDescriptors(td);
    int csdl_size = csdl.size();
    ArrayList<ConstantAction> newCongloms = new ArrayList<ConstantAction>();
    // we want to remove referenced primary/unique keys in the second
    // round.  This will ensure that self-referential constraints will
    // work OK.
    int tbr_size = 0;
    ConstraintDescriptor[] toBeRemoved = new ConstraintDescriptor[csdl_size];
    // let's go downwards, don't want to get messed up while removing
    for (int i = csdl_size - 1; i >= 0; i--) {
        ConstraintDescriptor cd = csdl.elementAt(i);
        int[] referencedColumns = cd.getReferencedColumns();
        int numRefCols = referencedColumns.length, j;
        boolean changed = false;
        for (j = 0; j < numRefCols; j++) {
            if (referencedColumns[j] > droppedColumnPosition)
                changed = true;
            if (referencedColumns[j] == droppedColumnPosition)
                break;
        }
        if (// column not referenced
        j == numRefCols) {
            if ((cd instanceof CheckConstraintDescriptor) && changed) {
                dd.dropConstraintDescriptor(cd, tc);
                for (j = 0; j < numRefCols; j++) {
                    if (referencedColumns[j] > droppedColumnPosition)
                        referencedColumns[j]--;
                }
                ((CheckConstraintDescriptor) cd).setReferencedColumnsDescriptor(new ReferencedColumnsDescriptorImpl(referencedColumns));
                dd.addConstraintDescriptor(cd, tc);
            }
            continue;
        }
        if (!cascade) {
            // 
            throw StandardException.newException(SQLState.LANG_PROVIDER_HAS_DEPENDENT_OBJECT, dm.getActionString(DependencyManager.DROP_COLUMN), columnName, "CONSTRAINT", cd.getConstraintName());
        }
        if (cd instanceof ReferencedKeyConstraintDescriptor) {
            // restrict will raise an error in invalidate if referenced
            toBeRemoved[tbr_size++] = cd;
            continue;
        }
        // drop now in all other cases
        dm.invalidateFor(cd, DependencyManager.DROP_CONSTRAINT, lcc);
        dropConstraint(cd, td, newCongloms, activation, lcc, true);
        activation.addWarning(StandardException.newWarning(SQLState.LANG_CONSTRAINT_DROPPED, cd.getConstraintName(), td.getName()));
    }
    for (int i = tbr_size - 1; i >= 0; i--) {
        ConstraintDescriptor cd = toBeRemoved[i];
        dropConstraint(cd, td, newCongloms, activation, lcc, false);
        activation.addWarning(StandardException.newWarning(SQLState.LANG_CONSTRAINT_DROPPED, cd.getConstraintName(), td.getName()));
        if (cascade) {
            ConstraintDescriptorList fkcdl = dd.getForeignKeys(cd.getUUID());
            for (ConstraintDescriptor fkcd : fkcdl) {
                dm.invalidateFor(fkcd, DependencyManager.DROP_CONSTRAINT, lcc);
                dropConstraint(fkcd, td, newCongloms, activation, lcc, true);
                activation.addWarning(StandardException.newWarning(SQLState.LANG_CONSTRAINT_DROPPED, fkcd.getConstraintName(), fkcd.getTableDescriptor().getName()));
            }
        }
        dm.invalidateFor(cd, DependencyManager.DROP_CONSTRAINT, lcc);
        dm.clearDependencies(lcc, cd);
    }
    /* If there are new backing conglomerates which must be
		 * created to replace a dropped shared conglomerate
		 * (where the shared conglomerate was dropped as part
		 * of a "drop constraint" call above), then create them
		 * now.  We do this *after* dropping all dependent
		 * constraints because we don't want to waste time
		 * creating a new conglomerate if it's just going to be
		 * dropped again as part of another "drop constraint".
		 */
    createNewBackingCongloms(newCongloms, (long[]) null);
    /*
         * The work we've done above, specifically the possible
         * dropping of primary key, foreign key, and unique constraints
         * and their underlying indexes, may have affected the table
         * descriptor. By re-reading the table descriptor here, we
         * ensure that the compressTable code is working with an
         * accurate table descriptor. Without this line, we may get
         * conglomerate-not-found errors and the like due to our
         * stale table descriptor.
         */
    td = dd.getTableDescriptor(tableId);
    compressTable();
    ColumnDescriptorList tab_cdl = td.getColumnDescriptorList();
    // drop the column from syscolumns
    dd.dropColumnDescriptor(td.getUUID(), columnName, tc);
    ColumnDescriptor[] cdlArray = new ColumnDescriptor[size - columnDescriptor.getPosition()];
    // 
    for (int i = columnDescriptor.getPosition(), j = 0; i < size; i++, j++) {
        ColumnDescriptor cd = tab_cdl.elementAt(i);
        dd.dropColumnDescriptor(td.getUUID(), cd.getColumnName(), tc);
        cd.setPosition(i);
        if (cd.isAutoincrement()) {
            cd.setAutoinc_create_or_modify_Start_Increment(ColumnDefinitionNode.CREATE_AUTOINCREMENT);
        }
        cdlArray[j] = cd;
    }
    dd.addDescriptorArray(cdlArray, td, DataDictionary.SYSCOLUMNS_CATALOG_NUM, false, tc);
    // By this time, the column has been removed from the table descriptor.
    // Now, go through all the triggers and regenerate their trigger action
    // SPS and rebind the generated trigger action sql. If the trigger
    // action is using the dropped column, it will get detected here. If
    // not, then we will have generated the internal trigger action sql
    // which matches the trigger action sql provided by the user.
    // 
    // eg of positive test case
    // create table atdc_16_tab1 (a1 integer, b1 integer, c1 integer);
    // create table atdc_16_tab2 (a2 integer, b2 integer, c2 integer);
    // create trigger atdc_16_trigger_1
    // after update of b1 on atdc_16_tab1
    // REFERENCING NEW AS newt
    // for each row
    // update atdc_16_tab2 set c2 = newt.c1
    // The internal representation for the trigger action before the column
    // is dropped is as follows
    // update atdc_16_tab2 set c2 =
    // org.apache.derby.iapi.db.Factory::getTriggerExecutionContext().
    // getONewRow().getInt(3)
    // After the drop column shown as below
    // alter table DERBY4998_SOFT_UPGRADE_RESTRICT drop column c11
    // The above internal representation of tigger action sql is not
    // correct anymore because column position of c1 in atdc_16_tab1 has
    // now changed from 3 to 2. Following while loop will regenerate it and
    // change it to as follows
    // update atdc_16_tab2 set c2 =
    // org.apache.derby.iapi.db.Factory::getTriggerExecutionContext().
    // getONewRow().getInt(2)
    // 
    // We could not do this before the actual column drop, because the
    // rebind would have still found the column being dropped in the
    // table descriptor and hence use of such a column in the trigger
    // action rebind would not have been caught.
    // For the table on which ALTER TABLE is getting performed, find out
    // all the SPSDescriptors that use that table as a provider. We are
    // looking for SPSDescriptors that have been created internally for
    // trigger action SPSes. Through those SPSDescriptors, we will be
    // able to get to the triggers dependent on the table being altered
    // Following will get all the dependent objects that are using
    // ALTER TABLE table as provider
    List<DependencyDescriptor> depsOnAlterTableList = dd.getProvidersDescriptorList(td.getObjectID().toString());
    for (DependencyDescriptor depOnAT : depsOnAlterTableList) {
        // Go through all the dependent objects on the table being altered
        DependableFinder dependent = depOnAT.getDependentFinder();
        // stored prepared statement.
        if (dependent.getSQLObjectType().equals(Dependable.STORED_PREPARED_STATEMENT)) {
            // Look for all the dependent objects that are using this
            // stored prepared statement as provider. We are only
            // interested in dependents that are triggers.
            List<DependencyDescriptor> depsTrigger = dd.getProvidersDescriptorList(depOnAT.getUUID().toString());
            for (DependencyDescriptor depsTriggerDesc : depsTrigger) {
                DependableFinder providerIsTrigger = depsTriggerDesc.getDependentFinder();
                // it is a trigger
                if (providerIsTrigger.getSQLObjectType().equals(Dependable.TRIGGER)) {
                    // Drop and recreate the trigger after regenerating
                    // it's trigger action plan. If the trigger action
                    // depends on the column being dropped, it will be
                    // caught here.
                    TriggerDescriptor trdToBeDropped = dd.getTriggerDescriptor(depsTriggerDesc.getUUID());
                    // First check for dependencies in the trigger's WHEN
                    // clause, if there is one.
                    UUID whenClauseId = trdToBeDropped.getWhenClauseId();
                    boolean gotDropped = false;
                    if (whenClauseId != null) {
                        gotDropped = columnDroppedAndTriggerDependencies(trdToBeDropped, whenClauseId, true, cascade, columnName);
                    }
                    // dependencies.
                    if (!gotDropped) {
                        columnDroppedAndTriggerDependencies(trdToBeDropped, trdToBeDropped.getActionId(), false, cascade, columnName);
                    }
                }
            }
        }
    }
    // Adjust the column permissions rows in SYSCOLPERMS to reflect the
    // changed column positions due to the dropped column:
    dd.updateSYSCOLPERMSforDropColumn(td.getUUID(), tc, columnDescriptor);
    // remove column descriptor from table descriptor. this fixes up the
    // list in case we were called recursively in order to cascade-drop a
    // dependent generated column.
    tab_cdl.remove(td.getColumnDescriptor(columnName));
}
Also used : DependencyDescriptor(org.apache.derby.iapi.sql.dictionary.DependencyDescriptor) ArrayList(java.util.ArrayList) ReferencedColumnsDescriptorImpl(org.apache.derby.catalog.types.ReferencedColumnsDescriptorImpl) CheckConstraintDescriptor(org.apache.derby.iapi.sql.dictionary.CheckConstraintDescriptor) DataDescriptorGenerator(org.apache.derby.iapi.sql.dictionary.DataDescriptorGenerator) ConstantAction(org.apache.derby.iapi.sql.execute.ConstantAction) DependableFinder(org.apache.derby.catalog.DependableFinder) ColumnDescriptorList(org.apache.derby.iapi.sql.dictionary.ColumnDescriptorList) FormatableBitSet(org.apache.derby.iapi.services.io.FormatableBitSet) UUID(org.apache.derby.catalog.UUID) ColumnDescriptor(org.apache.derby.iapi.sql.dictionary.ColumnDescriptor) ConstraintDescriptorList(org.apache.derby.iapi.sql.dictionary.ConstraintDescriptorList) TriggerDescriptor(org.apache.derby.iapi.sql.dictionary.TriggerDescriptor) CheckConstraintDescriptor(org.apache.derby.iapi.sql.dictionary.CheckConstraintDescriptor) ForeignKeyConstraintDescriptor(org.apache.derby.iapi.sql.dictionary.ForeignKeyConstraintDescriptor) ConstraintDescriptor(org.apache.derby.iapi.sql.dictionary.ConstraintDescriptor) ReferencedKeyConstraintDescriptor(org.apache.derby.iapi.sql.dictionary.ReferencedKeyConstraintDescriptor) ReferencedKeyConstraintDescriptor(org.apache.derby.iapi.sql.dictionary.ReferencedKeyConstraintDescriptor)

Example 4 with DependableFinder

use of org.apache.derby.catalog.DependableFinder in project derby by apache.

the class BasicDependencyManager method getDependencyDescriptorList.

/**
 * Turn a list of DependencyDescriptors into a list of Dependencies.
 *
 * @param storedList	The List of DependencyDescriptors representing
 *						stored dependencies.
 * @param providerForList The provider if this list is being created
 *                        for a list of dependents. Null otherwise.
 *
 * @return List		The converted List
 *
 * @exception StandardException thrown if something goes wrong
 */
private List<Dependency> getDependencyDescriptorList(List<DependencyDescriptor> storedList, Provider providerForList) throws StandardException {
    List<Dependency> retval = new ArrayList<Dependency>();
    if (!storedList.isEmpty()) {
        /* For each DependencyDescriptor, we need to instantiate
			 * object descriptors of the appropriate type for both
			 * the dependent and provider, create a Dependency with
			 * that Dependent and Provider and substitute the Dependency
			 * back into the same place in the List
			 * so that the call gets an enumerations of Dependencys.
			 */
        for (DependencyDescriptor depDesc : storedList) {
            Dependent tempD;
            Provider tempP;
            DependableFinder finder = depDesc.getDependentFinder();
            tempD = (Dependent) finder.getDependable(dd, depDesc.getUUID());
            if (providerForList != null) {
                // Use the provider being passed in.
                tempP = providerForList;
                // Sanity check the object identifiers match.
                if (SanityManager.DEBUG) {
                    if (!tempP.getObjectID().equals(depDesc.getProviderID())) {
                        SanityManager.THROWASSERT("mismatch providers");
                    }
                }
            } else {
                finder = depDesc.getProviderFinder();
                tempP = (Provider) finder.getDependable(dd, depDesc.getProviderID());
            }
            retval.add(new BasicDependency(tempD, tempP));
        }
    }
    return retval;
}
Also used : DependableFinder(org.apache.derby.catalog.DependableFinder) DependencyDescriptor(org.apache.derby.iapi.sql.dictionary.DependencyDescriptor) ArrayList(java.util.ArrayList) Dependency(org.apache.derby.iapi.sql.depend.Dependency) Dependent(org.apache.derby.iapi.sql.depend.Dependent) Provider(org.apache.derby.iapi.sql.depend.Provider)

Example 5 with DependableFinder

use of org.apache.derby.catalog.DependableFinder in project derby by apache.

the class SYSDEPENDSRowFactory method makeRow.

// ///////////////////////////////////////////////////////////////////////////
// 
// METHODS
// 
// ///////////////////////////////////////////////////////////////////////////
/**
 * Make a SYSDEPENDS row
 *
 * @param td DependencyDescriptor. If its null then we want to make an empty
 * row.
 *
 * @return	Row suitable for inserting into SYSDEPENDS.
 *
 * @exception   StandardException thrown on failure
 */
public ExecRow makeRow(TupleDescriptor td, TupleDescriptor parent) throws StandardException {
    DataValueDescriptor col;
    ExecRow row;
    String dependentID = null;
    DependableFinder dependentBloodhound = null;
    String providerID = null;
    DependableFinder providerBloodhound = null;
    if (td != null) {
        DependencyDescriptor dd = (DependencyDescriptor) td;
        dependentID = dd.getUUID().toString();
        dependentBloodhound = dd.getDependentFinder();
        if (dependentBloodhound == null) {
            throw StandardException.newException(SQLState.DEP_UNABLE_TO_STORE);
        }
        providerID = dd.getProviderID().toString();
        providerBloodhound = dd.getProviderFinder();
        if (providerBloodhound == null) {
            throw StandardException.newException(SQLState.DEP_UNABLE_TO_STORE);
        }
    }
    /* Insert info into sysdepends */
    /* RESOLVE - It would be nice to require less knowledge about sysdepends
		 * and have this be more table driven.
		 */
    /* Build the row to insert  */
    row = getExecutionFactory().getValueRow(SYSDEPENDS_COLUMN_COUNT);
    /* 1st column is DEPENDENTID (UUID - char(36)) */
    row.setColumn(SYSDEPENDS_DEPENDENTID, new SQLChar(dependentID));
    /* 2nd column is DEPENDENTFINDER */
    row.setColumn(SYSDEPENDS_DEPENDENTTYPE, new UserType(dependentBloodhound));
    /* 3rd column is PROVIDERID (UUID - char(36)) */
    row.setColumn(SYSDEPENDS_PROVIDERID, new SQLChar(providerID));
    /* 4th column is PROVIDERFINDER */
    row.setColumn(SYSDEPENDS_PROVIDERTYPE, new UserType(providerBloodhound));
    return row;
}
Also used : DependableFinder(org.apache.derby.catalog.DependableFinder) DependencyDescriptor(org.apache.derby.iapi.sql.dictionary.DependencyDescriptor) ExecRow(org.apache.derby.iapi.sql.execute.ExecRow) SQLChar(org.apache.derby.iapi.types.SQLChar) DataValueDescriptor(org.apache.derby.iapi.types.DataValueDescriptor) UserType(org.apache.derby.iapi.types.UserType)

Aggregations

DependableFinder (org.apache.derby.catalog.DependableFinder)6 DependencyDescriptor (org.apache.derby.iapi.sql.dictionary.DependencyDescriptor)4 UUID (org.apache.derby.catalog.UUID)3 ArrayList (java.util.ArrayList)2 Provider (org.apache.derby.iapi.sql.depend.Provider)2 ColumnDescriptor (org.apache.derby.iapi.sql.dictionary.ColumnDescriptor)2 DataValueDescriptor (org.apache.derby.iapi.types.DataValueDescriptor)2 Statement (java.sql.Statement)1 ReferencedColumnsDescriptorImpl (org.apache.derby.catalog.types.ReferencedColumnsDescriptorImpl)1 ContextManager (org.apache.derby.iapi.services.context.ContextManager)1 FormatableBitSet (org.apache.derby.iapi.services.io.FormatableBitSet)1 Dependency (org.apache.derby.iapi.sql.depend.Dependency)1 DependencyManager (org.apache.derby.iapi.sql.depend.DependencyManager)1 Dependent (org.apache.derby.iapi.sql.depend.Dependent)1 ProviderInfo (org.apache.derby.iapi.sql.depend.ProviderInfo)1 CheckConstraintDescriptor (org.apache.derby.iapi.sql.dictionary.CheckConstraintDescriptor)1 ColumnDescriptorList (org.apache.derby.iapi.sql.dictionary.ColumnDescriptorList)1 ConstraintDescriptor (org.apache.derby.iapi.sql.dictionary.ConstraintDescriptor)1 ConstraintDescriptorList (org.apache.derby.iapi.sql.dictionary.ConstraintDescriptorList)1 DataDescriptorGenerator (org.apache.derby.iapi.sql.dictionary.DataDescriptorGenerator)1