session/
context.rs

1// Copyright 2023 Greptime Team
2//
3// Licensed under the Apache License, Version 2.0 (the "License");
4// you may not use this file except in compliance with the License.
5// You may obtain a copy of the License at
6//
7//     http://www.apache.org/licenses/LICENSE-2.0
8//
9// Unless required by applicable law or agreed to in writing, software
10// distributed under the License is distributed on an "AS IS" BASIS,
11// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12// See the License for the specific language governing permissions and
13// limitations under the License.
14
15use std::collections::HashMap;
16use std::fmt::{Display, Formatter};
17use std::net::SocketAddr;
18use std::sync::{Arc, RwLock};
19use std::time::Duration;
20
21use api::v1::ExplainOptions;
22use api::v1::region::RegionRequestHeader;
23use arc_swap::ArcSwap;
24use auth::UserInfoRef;
25use common_catalog::consts::{DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME};
26use common_catalog::{build_db_string, parse_catalog_and_schema_from_db_string};
27use common_recordbatch::cursor::RecordBatchStreamCursor;
28use common_telemetry::warn;
29use common_time::Timezone;
30use common_time::timezone::parse_timezone;
31use datafusion_common::config::ConfigOptions;
32use derive_builder::Builder;
33use sql::dialect::{Dialect, GenericDialect, GreptimeDbDialect, MySqlDialect, PostgreSqlDialect};
34
35use crate::protocol_ctx::ProtocolCtx;
36use crate::session_config::{PGByteaOutputValue, PGDateOrder, PGDateTimeStyle};
37use crate::{MutableInner, ReadPreference};
38
39pub type QueryContextRef = Arc<QueryContext>;
40pub type ConnInfoRef = Arc<ConnInfo>;
41
42const CURSOR_COUNT_WARNING_LIMIT: usize = 10;
43
44#[derive(Debug, Builder, Clone)]
45#[builder(pattern = "owned")]
46#[builder(build_fn(skip))]
47pub struct QueryContext {
48    current_catalog: String,
49    /// mapping of RegionId to SequenceNumber, for snapshot read, meaning that the read should only
50    /// container data that was committed before(and include) the given sequence number
51    /// this field will only be filled if extensions contains a pair of "snapshot_read" and "true"
52    snapshot_seqs: Arc<RwLock<HashMap<u64, u64>>>,
53    /// Mappings of the RegionId to the minimal sequence of SST file to scan.
54    sst_min_sequences: Arc<RwLock<HashMap<u64, u64>>>,
55    // we use Arc<RwLock>> for modifiable fields
56    #[builder(default)]
57    mutable_session_data: Arc<RwLock<MutableInner>>,
58    #[builder(default)]
59    mutable_query_context_data: Arc<RwLock<QueryContextMutableFields>>,
60    sql_dialect: Arc<dyn Dialect + Send + Sync>,
61    #[builder(default)]
62    extensions: HashMap<String, String>,
63    /// The configuration parameter are used to store the parameters that are set by the user
64    #[builder(default)]
65    configuration_parameter: Arc<ConfigurationVariables>,
66    /// Track which protocol the query comes from.
67    #[builder(default)]
68    channel: Channel,
69    /// Process id for managing on-going queries
70    #[builder(default)]
71    process_id: u32,
72    /// Connection information
73    #[builder(default)]
74    conn_info: ConnInfo,
75    /// Protocol specific context
76    #[builder(default)]
77    protocol_ctx: ProtocolCtx,
78}
79
80/// This fields hold data that is only valid to current query context
81#[derive(Debug, Builder, Clone, Default)]
82pub struct QueryContextMutableFields {
83    warning: Option<String>,
84    // TODO: remove this when format is supported in datafusion
85    explain_format: Option<String>,
86    /// Explain options to control the verbose analyze output.
87    explain_options: Option<ExplainOptions>,
88}
89
90impl Display for QueryContext {
91    fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result {
92        write!(
93            f,
94            "QueryContext{{catalog: {}, schema: {}}}",
95            self.current_catalog(),
96            self.current_schema()
97        )
98    }
99}
100
101impl QueryContextBuilder {
102    pub fn current_schema(mut self, schema: String) -> Self {
103        if self.mutable_session_data.is_none() {
104            self.mutable_session_data = Some(Arc::new(RwLock::new(MutableInner::default())));
105        }
106
107        // safe for unwrap because previous none check
108        self.mutable_session_data
109            .as_mut()
110            .unwrap()
111            .write()
112            .unwrap()
113            .schema = schema;
114        self
115    }
116
117    pub fn timezone(mut self, timezone: Timezone) -> Self {
118        if self.mutable_session_data.is_none() {
119            self.mutable_session_data = Some(Arc::new(RwLock::new(MutableInner::default())));
120        }
121
122        self.mutable_session_data
123            .as_mut()
124            .unwrap()
125            .write()
126            .unwrap()
127            .timezone = timezone;
128        self
129    }
130
131    pub fn explain_options(mut self, explain_options: Option<ExplainOptions>) -> Self {
132        self.mutable_query_context_data
133            .get_or_insert_default()
134            .write()
135            .unwrap()
136            .explain_options = explain_options;
137        self
138    }
139
140    pub fn read_preference(mut self, read_preference: ReadPreference) -> Self {
141        self.mutable_session_data
142            .get_or_insert_default()
143            .write()
144            .unwrap()
145            .read_preference = read_preference;
146        self
147    }
148}
149
150impl From<&RegionRequestHeader> for QueryContext {
151    fn from(value: &RegionRequestHeader) -> Self {
152        if let Some(ctx) = &value.query_context {
153            ctx.clone().into()
154        } else {
155            QueryContextBuilder::default().build()
156        }
157    }
158}
159
160impl From<api::v1::QueryContext> for QueryContext {
161    fn from(ctx: api::v1::QueryContext) -> Self {
162        let sequences = ctx.snapshot_seqs.as_ref();
163        QueryContextBuilder::default()
164            .current_catalog(ctx.current_catalog)
165            .current_schema(ctx.current_schema)
166            .timezone(parse_timezone(Some(&ctx.timezone)))
167            .extensions(ctx.extensions)
168            .channel(ctx.channel.into())
169            .snapshot_seqs(Arc::new(RwLock::new(
170                sequences
171                    .map(|x| x.snapshot_seqs.clone())
172                    .unwrap_or_default(),
173            )))
174            .sst_min_sequences(Arc::new(RwLock::new(
175                sequences
176                    .map(|x| x.sst_min_sequences.clone())
177                    .unwrap_or_default(),
178            )))
179            .explain_options(ctx.explain)
180            .build()
181    }
182}
183
184impl From<QueryContext> for api::v1::QueryContext {
185    fn from(
186        QueryContext {
187            current_catalog,
188            mutable_session_data: mutable_inner,
189            extensions,
190            channel,
191            snapshot_seqs,
192            sst_min_sequences,
193            mutable_query_context_data,
194            ..
195        }: QueryContext,
196    ) -> Self {
197        let explain = mutable_query_context_data.read().unwrap().explain_options;
198        let mutable_inner = mutable_inner.read().unwrap();
199        api::v1::QueryContext {
200            current_catalog,
201            current_schema: mutable_inner.schema.clone(),
202            timezone: mutable_inner.timezone.to_string(),
203            extensions,
204            channel: channel as u32,
205            snapshot_seqs: Some(api::v1::SnapshotSequences {
206                snapshot_seqs: snapshot_seqs.read().unwrap().clone(),
207                sst_min_sequences: sst_min_sequences.read().unwrap().clone(),
208            }),
209            explain,
210        }
211    }
212}
213
214impl From<&QueryContext> for api::v1::QueryContext {
215    fn from(ctx: &QueryContext) -> Self {
216        ctx.clone().into()
217    }
218}
219
220impl QueryContext {
221    pub fn arc() -> QueryContextRef {
222        Arc::new(QueryContextBuilder::default().build())
223    }
224
225    /// Create a new  datafusion's ConfigOptions instance based on the current QueryContext.
226    pub fn create_config_options(&self) -> ConfigOptions {
227        let mut config = ConfigOptions::default();
228        config.execution.time_zone = self.timezone().to_string();
229        config
230    }
231
232    pub fn with(catalog: &str, schema: &str) -> QueryContext {
233        QueryContextBuilder::default()
234            .current_catalog(catalog.to_string())
235            .current_schema(schema.to_string())
236            .build()
237    }
238
239    pub fn with_channel(catalog: &str, schema: &str, channel: Channel) -> QueryContext {
240        QueryContextBuilder::default()
241            .current_catalog(catalog.to_string())
242            .current_schema(schema.to_string())
243            .channel(channel)
244            .build()
245    }
246
247    pub fn with_db_name(db_name: Option<&str>) -> QueryContext {
248        let (catalog, schema) = db_name
249            .map(|db| {
250                let (catalog, schema) = parse_catalog_and_schema_from_db_string(db);
251                (catalog, schema)
252            })
253            .unwrap_or_else(|| {
254                (
255                    DEFAULT_CATALOG_NAME.to_string(),
256                    DEFAULT_SCHEMA_NAME.to_string(),
257                )
258            });
259        QueryContextBuilder::default()
260            .current_catalog(catalog)
261            .current_schema(schema.to_string())
262            .build()
263    }
264
265    pub fn current_schema(&self) -> String {
266        self.mutable_session_data.read().unwrap().schema.clone()
267    }
268
269    pub fn set_current_schema(&self, new_schema: &str) {
270        self.mutable_session_data.write().unwrap().schema = new_schema.to_string();
271    }
272
273    pub fn current_catalog(&self) -> &str {
274        &self.current_catalog
275    }
276
277    pub fn set_current_catalog(&mut self, new_catalog: &str) {
278        self.current_catalog = new_catalog.to_string();
279    }
280
281    pub fn sql_dialect(&self) -> &(dyn Dialect + Send + Sync) {
282        &*self.sql_dialect
283    }
284
285    pub fn get_db_string(&self) -> String {
286        let catalog = self.current_catalog();
287        let schema = self.current_schema();
288        build_db_string(catalog, &schema)
289    }
290
291    pub fn timezone(&self) -> Timezone {
292        self.mutable_session_data.read().unwrap().timezone.clone()
293    }
294
295    pub fn set_timezone(&self, timezone: Timezone) {
296        self.mutable_session_data.write().unwrap().timezone = timezone;
297    }
298
299    pub fn read_preference(&self) -> ReadPreference {
300        self.mutable_session_data.read().unwrap().read_preference
301    }
302
303    pub fn set_read_preference(&self, read_preference: ReadPreference) {
304        self.mutable_session_data.write().unwrap().read_preference = read_preference;
305    }
306
307    pub fn current_user(&self) -> UserInfoRef {
308        self.mutable_session_data.read().unwrap().user_info.clone()
309    }
310
311    pub fn set_current_user(&self, user: UserInfoRef) {
312        self.mutable_session_data.write().unwrap().user_info = user;
313    }
314
315    pub fn set_extension<S1: Into<String>, S2: Into<String>>(&mut self, key: S1, value: S2) {
316        self.extensions.insert(key.into(), value.into());
317    }
318
319    pub fn extension<S: AsRef<str>>(&self, key: S) -> Option<&str> {
320        self.extensions.get(key.as_ref()).map(|v| v.as_str())
321    }
322
323    pub fn extensions(&self) -> HashMap<String, String> {
324        self.extensions.clone()
325    }
326
327    /// Default to double quote and fallback to back quote
328    pub fn quote_style(&self) -> char {
329        if self.sql_dialect().is_delimited_identifier_start('"') {
330            '"'
331        } else if self.sql_dialect().is_delimited_identifier_start('\'') {
332            '\''
333        } else {
334            '`'
335        }
336    }
337
338    pub fn configuration_parameter(&self) -> &ConfigurationVariables {
339        &self.configuration_parameter
340    }
341
342    pub fn channel(&self) -> Channel {
343        self.channel
344    }
345
346    pub fn set_channel(&mut self, channel: Channel) {
347        self.channel = channel;
348    }
349
350    pub fn warning(&self) -> Option<String> {
351        self.mutable_query_context_data
352            .read()
353            .unwrap()
354            .warning
355            .clone()
356    }
357
358    pub fn set_warning(&self, msg: String) {
359        self.mutable_query_context_data.write().unwrap().warning = Some(msg);
360    }
361
362    pub fn explain_format(&self) -> Option<String> {
363        self.mutable_query_context_data
364            .read()
365            .unwrap()
366            .explain_format
367            .clone()
368    }
369
370    pub fn set_explain_format(&self, format: String) {
371        self.mutable_query_context_data
372            .write()
373            .unwrap()
374            .explain_format = Some(format);
375    }
376
377    pub fn explain_verbose(&self) -> bool {
378        self.mutable_query_context_data
379            .read()
380            .unwrap()
381            .explain_options
382            .map(|opts| opts.verbose)
383            .unwrap_or(false)
384    }
385
386    pub fn set_explain_verbose(&self, verbose: bool) {
387        self.mutable_query_context_data
388            .write()
389            .unwrap()
390            .explain_options
391            .get_or_insert_default()
392            .verbose = verbose;
393    }
394
395    pub fn query_timeout(&self) -> Option<Duration> {
396        self.mutable_session_data.read().unwrap().query_timeout
397    }
398
399    pub fn query_timeout_as_millis(&self) -> u128 {
400        let timeout = self.mutable_session_data.read().unwrap().query_timeout;
401        if let Some(t) = timeout {
402            return t.as_millis();
403        }
404        0
405    }
406
407    pub fn set_query_timeout(&self, timeout: Duration) {
408        self.mutable_session_data.write().unwrap().query_timeout = Some(timeout);
409    }
410
411    pub fn insert_cursor(&self, name: String, rb: RecordBatchStreamCursor) {
412        let mut guard = self.mutable_session_data.write().unwrap();
413        guard.cursors.insert(name, Arc::new(rb));
414
415        let cursor_count = guard.cursors.len();
416        if cursor_count > CURSOR_COUNT_WARNING_LIMIT {
417            warn!("Current connection has {} open cursors", cursor_count);
418        }
419    }
420
421    pub fn remove_cursor(&self, name: &str) {
422        let mut guard = self.mutable_session_data.write().unwrap();
423        guard.cursors.remove(name);
424    }
425
426    pub fn get_cursor(&self, name: &str) -> Option<Arc<RecordBatchStreamCursor>> {
427        let guard = self.mutable_session_data.read().unwrap();
428        let rb = guard.cursors.get(name);
429        rb.cloned()
430    }
431
432    pub fn snapshots(&self) -> HashMap<u64, u64> {
433        self.snapshot_seqs.read().unwrap().clone()
434    }
435
436    pub fn get_snapshot(&self, region_id: u64) -> Option<u64> {
437        self.snapshot_seqs.read().unwrap().get(&region_id).cloned()
438    }
439
440    /// Returns `true` if the session can cast strings to numbers in MySQL style.
441    pub fn auto_string_to_numeric(&self) -> bool {
442        matches!(self.channel, Channel::Mysql)
443    }
444
445    /// Finds the minimal sequence of SST files to scan of a Region.
446    pub fn sst_min_sequence(&self, region_id: u64) -> Option<u64> {
447        self.sst_min_sequences
448            .read()
449            .unwrap()
450            .get(&region_id)
451            .copied()
452    }
453
454    pub fn process_id(&self) -> u32 {
455        self.process_id
456    }
457
458    /// Get client information
459    pub fn conn_info(&self) -> &ConnInfo {
460        &self.conn_info
461    }
462
463    pub fn protocol_ctx(&self) -> &ProtocolCtx {
464        &self.protocol_ctx
465    }
466
467    pub fn set_protocol_ctx(&mut self, protocol_ctx: ProtocolCtx) {
468        self.protocol_ctx = protocol_ctx;
469    }
470}
471
472impl QueryContextBuilder {
473    pub fn build(self) -> QueryContext {
474        let channel = self.channel.unwrap_or_default();
475        QueryContext {
476            current_catalog: self
477                .current_catalog
478                .unwrap_or_else(|| DEFAULT_CATALOG_NAME.to_string()),
479            snapshot_seqs: self.snapshot_seqs.unwrap_or_default(),
480            sst_min_sequences: self.sst_min_sequences.unwrap_or_default(),
481            mutable_session_data: self.mutable_session_data.unwrap_or_default(),
482            mutable_query_context_data: self.mutable_query_context_data.unwrap_or_default(),
483            sql_dialect: self
484                .sql_dialect
485                .unwrap_or_else(|| Arc::new(GreptimeDbDialect {})),
486            extensions: self.extensions.unwrap_or_default(),
487            configuration_parameter: self
488                .configuration_parameter
489                .unwrap_or_else(|| Arc::new(ConfigurationVariables::default())),
490            channel,
491            process_id: self.process_id.unwrap_or_default(),
492            conn_info: self.conn_info.unwrap_or_default(),
493            protocol_ctx: self.protocol_ctx.unwrap_or_default(),
494        }
495    }
496
497    pub fn set_extension(mut self, key: String, value: String) -> Self {
498        self.extensions
499            .get_or_insert_with(HashMap::new)
500            .insert(key, value);
501        self
502    }
503}
504
505#[derive(Debug, Clone, Default)]
506pub struct ConnInfo {
507    pub client_addr: Option<SocketAddr>,
508    pub channel: Channel,
509}
510
511impl Display for ConnInfo {
512    fn fmt(&self, f: &mut Formatter) -> std::fmt::Result {
513        write!(
514            f,
515            "{}[{}]",
516            self.channel,
517            self.client_addr
518                .map(|addr| addr.to_string())
519                .as_deref()
520                .unwrap_or("unknown client addr")
521        )
522    }
523}
524
525impl ConnInfo {
526    pub fn new(client_addr: Option<SocketAddr>, channel: Channel) -> Self {
527        Self {
528            client_addr,
529            channel,
530        }
531    }
532}
533
534#[derive(Debug, PartialEq, Default, Clone, Copy)]
535#[repr(u8)]
536pub enum Channel {
537    #[default]
538    Unknown = 0,
539
540    Mysql = 1,
541    Postgres = 2,
542    HttpSql = 3,
543    Prometheus = 4,
544    Otlp = 5,
545    Grpc = 6,
546    Influx = 7,
547    Opentsdb = 8,
548    Loki = 9,
549    Elasticsearch = 10,
550    Jaeger = 11,
551    Log = 12,
552    Promql = 13,
553}
554
555impl From<u32> for Channel {
556    fn from(value: u32) -> Self {
557        match value {
558            1 => Self::Mysql,
559            2 => Self::Postgres,
560            3 => Self::HttpSql,
561            4 => Self::Prometheus,
562            5 => Self::Otlp,
563            6 => Self::Grpc,
564            7 => Self::Influx,
565            8 => Self::Opentsdb,
566            9 => Self::Loki,
567            10 => Self::Elasticsearch,
568            11 => Self::Jaeger,
569            12 => Self::Log,
570            13 => Self::Promql,
571            _ => Self::Unknown,
572        }
573    }
574}
575
576impl Channel {
577    pub fn dialect(&self) -> Arc<dyn Dialect + Send + Sync> {
578        match self {
579            Channel::Mysql => Arc::new(MySqlDialect {}),
580            Channel::Postgres => Arc::new(PostgreSqlDialect {}),
581            _ => Arc::new(GenericDialect {}),
582        }
583    }
584}
585
586impl Display for Channel {
587    fn fmt(&self, f: &mut Formatter) -> std::fmt::Result {
588        write!(f, "{}", self.as_ref())
589    }
590}
591
592impl AsRef<str> for Channel {
593    fn as_ref(&self) -> &str {
594        match self {
595            Channel::Mysql => "mysql",
596            Channel::Postgres => "postgres",
597            Channel::HttpSql => "httpsql",
598            Channel::Prometheus => "prometheus",
599            Channel::Otlp => "otlp",
600            Channel::Grpc => "grpc",
601            Channel::Influx => "influx",
602            Channel::Opentsdb => "opentsdb",
603            Channel::Loki => "loki",
604            Channel::Elasticsearch => "elasticsearch",
605            Channel::Jaeger => "jaeger",
606            Channel::Log => "log",
607            Channel::Promql => "promql",
608            Channel::Unknown => "unknown",
609        }
610    }
611}
612
613#[derive(Default, Debug)]
614pub struct ConfigurationVariables {
615    postgres_bytea_output: ArcSwap<PGByteaOutputValue>,
616    pg_datestyle_format: ArcSwap<(PGDateTimeStyle, PGDateOrder)>,
617    allow_query_fallback: ArcSwap<bool>,
618}
619
620impl Clone for ConfigurationVariables {
621    fn clone(&self) -> Self {
622        Self {
623            postgres_bytea_output: ArcSwap::new(self.postgres_bytea_output.load().clone()),
624            pg_datestyle_format: ArcSwap::new(self.pg_datestyle_format.load().clone()),
625            allow_query_fallback: ArcSwap::new(self.allow_query_fallback.load().clone()),
626        }
627    }
628}
629
630impl ConfigurationVariables {
631    pub fn new() -> Self {
632        Self::default()
633    }
634
635    pub fn set_postgres_bytea_output(&self, value: PGByteaOutputValue) {
636        let _ = self.postgres_bytea_output.swap(Arc::new(value));
637    }
638
639    pub fn postgres_bytea_output(&self) -> Arc<PGByteaOutputValue> {
640        self.postgres_bytea_output.load().clone()
641    }
642
643    pub fn pg_datetime_style(&self) -> Arc<(PGDateTimeStyle, PGDateOrder)> {
644        self.pg_datestyle_format.load().clone()
645    }
646
647    pub fn set_pg_datetime_style(&self, style: PGDateTimeStyle, order: PGDateOrder) {
648        self.pg_datestyle_format.swap(Arc::new((style, order)));
649    }
650
651    pub fn allow_query_fallback(&self) -> bool {
652        **self.allow_query_fallback.load()
653    }
654
655    pub fn set_allow_query_fallback(&self, allow: bool) {
656        self.allow_query_fallback.swap(Arc::new(allow));
657    }
658}
659
660#[cfg(test)]
661mod test {
662    use common_catalog::consts::DEFAULT_CATALOG_NAME;
663
664    use super::*;
665    use crate::Session;
666    use crate::context::Channel;
667
668    #[test]
669    fn test_session() {
670        let session = Session::new(
671            Some("127.0.0.1:9000".parse().unwrap()),
672            Channel::Mysql,
673            Default::default(),
674            100,
675        );
676        // test user_info
677        assert_eq!(session.user_info().username(), "greptime");
678
679        // test channel
680        assert_eq!(session.conn_info().channel, Channel::Mysql);
681        let client_addr = session.conn_info().client_addr.as_ref().unwrap();
682        assert_eq!(client_addr.ip().to_string(), "127.0.0.1");
683        assert_eq!(client_addr.port(), 9000);
684
685        assert_eq!("mysql[127.0.0.1:9000]", session.conn_info().to_string());
686        assert_eq!(100, session.process_id());
687    }
688
689    #[test]
690    fn test_context_db_string() {
691        let context = QueryContext::with("a0b1c2d3", "test");
692        assert_eq!("a0b1c2d3-test", context.get_db_string());
693
694        let context = QueryContext::with(DEFAULT_CATALOG_NAME, "test");
695        assert_eq!("test", context.get_db_string());
696    }
697}
session/context.rs

session/
context.rs
