in parquet-cli/src/main/java/org/apache/parquet/cli/commands/ShowPagesCommand.java [78:150]
public int run() throws IOException {
Preconditions.checkArgument(targets != null && targets.size() >= 1, "A Parquet file is required.");
Preconditions.checkArgument(targets.size() == 1, "Cannot process multiple Parquet files.");
// Even though the implementation is separated the functionality is logically related so placed in the same
// command.
if (raw) {
try (RawPagesReader reader =
new RawPagesReader(HadoopInputFile.fromPath(qualifiedPath(targets.get(0)), getConf()), columns)) {
reader.listPages(console);
}
return 0;
}
String source = targets.get(0);
try (ParquetFileReader reader = ParquetFileReader.open(getConf(), qualifiedPath(source))) {
MessageType schema = reader.getFileMetaData().getSchema();
Map<ColumnDescriptor, PrimitiveType> columns = Maps.newLinkedHashMap();
if (this.columns == null || this.columns.isEmpty()) {
for (ColumnDescriptor descriptor : schema.getColumns()) {
columns.put(descriptor, primitive(schema, descriptor.getPath()));
}
} else {
for (String column : this.columns) {
columns.put(descriptor(column, schema), primitive(column, schema));
}
}
CompressionCodecName codec =
reader.getRowGroups().get(0).getColumns().get(0).getCodec();
// accumulate formatted lines to print by column
Map<String, List<String>> formatted = Maps.newLinkedHashMap();
PageFormatter formatter = new PageFormatter();
PageReadStore pageStore;
int rowGroupNum = 0;
while ((pageStore = reader.readNextRowGroup()) != null) {
for (ColumnDescriptor descriptor : columns.keySet()) {
List<String> lines = formatted.get(columnName(descriptor));
if (lines == null) {
lines = Lists.newArrayList();
formatted.put(columnName(descriptor), lines);
}
formatter.setContext(rowGroupNum, columns.get(descriptor), codec);
PageReader pages = pageStore.getPageReader(descriptor);
DictionaryPage dict = pages.readDictionaryPage();
if (dict != null) {
lines.add(formatter.format(dict));
}
DataPage page;
while ((page = pages.readPage()) != null) {
lines.add(formatter.format(page));
}
}
rowGroupNum += 1;
pageStore.close();
}
// TODO: Show total column size and overall size per value in the column summary line
for (String columnName : formatted.keySet()) {
console.info(String.format(
"\nColumn: %s\n%s", columnName, new TextStringBuilder(80).appendPadding(80, '-')));
console.info(formatter.getHeader());
for (String line : formatted.get(columnName)) {
console.info(line);
}
console.info("");
}
}
return 0;
}