Materialize a single data source
(self, source: str)
| 71 | return timings |
| 72 | |
| 73 | def _materialize_source(self, source: str): |
| 74 | """Materialize a single data source""" |
| 75 | table_name = self.TABLE_NAMES[source] |
| 76 | |
| 77 | # Drop existing table if it exists |
| 78 | self.conn.execute(f"DROP TABLE IF EXISTS {table_name}") |
| 79 | |
| 80 | if source == 'samples': |
| 81 | # Create enriched samples table with computed columns |
| 82 | query = f""" |
| 83 | CREATE TABLE {table_name} AS |
| 84 | SELECT |
| 85 | samples.*, |
| 86 | -- Computed columns |
| 87 | COALESCE(REGEXP_REPLACE(FILENAME, '[0-9]+', '*', 'g'), '-') AS FILENAMESUM, |
| 88 | CASE |
| 89 | WHEN FILENAME IS NULL THEN '-' |
| 90 | WHEN REGEXP_MATCHES(FILENAME, '\\.([^\\.]+)$') |
| 91 | THEN REGEXP_EXTRACT(FILENAME, '(\\.([^\\.]+))$', 1) |
| 92 | ELSE '-' |
| 93 | END AS FEXT, |
| 94 | CASE |
| 95 | WHEN COMM LIKE 'ora_p%' |
| 96 | THEN regexp_replace(COMM, '(?:p[0-9a-z]+_)', 'p*_', 'g') |
| 97 | ELSE regexp_replace(COMM, '[0-9]+', '*', 'g') |
| 98 | END AS COMM2, |
| 99 | CASE |
| 100 | WHEN EXTRA_INFO LIKE '%"connection"%' |
| 101 | THEN json_extract_string(EXTRA_INFO, '$.connection') |
| 102 | ELSE '-' |
| 103 | END AS CONNECTION |
| 104 | FROM read_csv_auto('{self.datadir}/xcapture_samples_*.csv') AS samples |
| 105 | """ |
| 106 | elif source == 'syscend': |
| 107 | query = f""" |
| 108 | CREATE TABLE {table_name} AS |
| 109 | SELECT * FROM read_csv_auto('{self.datadir}/xcapture_syscend_*.csv') |
| 110 | """ |
| 111 | elif source == 'iorqend': |
| 112 | query = f""" |
| 113 | CREATE TABLE {table_name} AS |
| 114 | SELECT * FROM read_csv_auto('{self.datadir}/xcapture_iorqend_*.csv') |
| 115 | """ |
| 116 | elif source == 'kstacks': |
| 117 | query = f""" |
| 118 | CREATE TABLE {table_name} AS |
| 119 | SELECT |
| 120 | KSTACK_HASH, |
| 121 | KSTACK_SYMS |
| 122 | FROM read_csv_auto('{self.datadir}/xcapture_kstacks_*.csv') |
| 123 | """ |
| 124 | elif source == 'ustacks': |
| 125 | query = f""" |
| 126 | CREATE TABLE {table_name} AS |
| 127 | SELECT |
| 128 | USTACK_HASH, |
| 129 | USTACK_SYMS |
| 130 | FROM read_csv_auto('{self.datadir}/xcapture_ustacks_*.csv') |
no test coverage detected