我需要搜索结果为"Achieved"的记录。该数据是一个JSON对象，位于Postgres 9.4中的text列中。
我尝试使用：
notes.data similar to '%"Legal"%' AND
notes.data similar to '%"Achieved"%' AND
notes.data similar to '%"Education"%' AND
notes.data similar to '%"Achieved"%'
但是，单词"Achieved"的实例不一定与"Legal"在同一位置匹配。正如您在下面的示例数据中所看到的，在Goals Array部分中，可能会有进展更新 - ##评论是我的说明。 （您需要删除那些才能获得有效的JSON格式！）
查询会将状态"Legal"拉为正常，但第二个条件"Achieved"可能与最新状态更新不匹配。
我已经阅读了JSON中的文本查询，但我对此结构感到茫然。我把它放在jsonlint.com中，它验证为正确的JSON。有人可以帮忙吗？
以下是完整的查询，后跟数据样本：
SELECT
clients.name_lastfirst_cs,
clients.client_id,
notes.date_service,
services.code,
services.name,
notes.data,
notes.zrud_template,
staff.staff_name_cs
FROM
public.clients,
public.staff,
public.notes,
public.services
WHERE
clients.zzud_client = notes.zrud_client AND
notes.zrud_staff = staff.zzud_staff AND
notes.zrud_service = services.zzud_service AND
notes.data similar to '%Legal%' AND notes.data similar to '%Achieved%' AND
notes.data similar to '%Substance Abuse%' AND notes.data similar to '%Achieved%'AND
services.code = '000502' AND
notes.date_service BETWEEN '07/01/2014' AND '04/15/2015'
ORDER BY clients.name_lastfirst_cs ASC;
notes.data中的示例值：
{
"DxArray": [
[
[
"Axis I",
"305.20",
"Cannabis Abuse ",
1,
"4F9E9DC8-D2ED-433A-A129-5696B34A866C"
]
],
[
[
"Axis I",
"304.20",
"Cocaine Dependence In Full Remission ",
2,
"37BA3F7C-B376-4DFD-82BE-04F26E3A9F2A"
]
],
[
[
"Axis II",
"799.9",
"Deferred",
3,
"DF80F337-DEF6-4265-9CBA-0F4ECA3E1A4D"
]
],
[
[
"Axis III",
null,
"Chronic Pain per client report. ",
4,
"3C49E6EC-9C95-4CA7-8005-DF70D071F5F1"
]
],
[
[
"Axis IV",
null,
"legal issues ",
5,
"C3114B75-ECDB-40AB-AD37-544D341549BA"
]
],
[
[
"Axis V",
null,
"GAF=60",
6,
"3C1225B3-1B5B-4F4C-A75A-1A2E2F46AAC2"
]
],
[
[
"Axis",
"Code",
"Diagnosis"
]
]
],
"update": "03/19/2015", ## date of update
"": "",
"header_HTML": "",
"Barriers": "Violation ",
"DOB": "01/15/1980",
"previewBounds": "",
"p0": 27,
"zrud_template": "82552FEB-8408-4A2D-81CF-564CC04108F8",
"fv_location": "Office/Agency",
"zzud_client": "07863B38-3793-49B5-A3C4-CB2186AAA48E",
"name_first": "Roger",
"Date": "03/10/2014", ## date of service plan
"Preferences": "\"I want to get finished with probation.",
"ID": "RABR010170",
"gender": 1,
"fv_servicecode": "000502",
"zrud_service": "D303BD9A-FA08-45DB-8CB4-BCE68FF83AE6",
"zzud_staff": "884AC915-0D3A-4826-B248-11498323DA21",
"fv_servicename": "Service Plan",
"axis_HTML": "",
"goals_HTML": "",
"zrud_link_service": "6171D722-632E-4BEC-B521-F598C1888040",
"is_cloned": "F8158F4B-F269-4BC9-AA20-48872F59C2E4",
"previous_plans": {
"1": {},
"2": {},
"3": {},
"4": {},
"5": {},
"6": {},
"7": {}
},
"GoalsArray": [ ## beginning of goals
[
[
"Legal", ## goal 1
" He will not engage in criminal activity.",
" Comply with all requirements of probation. No new crime or probation violation. ",
" Frequency: At least once per month and as needed.",
" Compliance with Care Manager Plan and all probation requirements.",
"Ted Therapist, CM\n\nClient\n\nProbation Officer ",
99,
"03/10/2014" ## date of goal creation
],
[
[
"09/06/2014", ## new target date
"03/19/2015", ## date updated
"Ongoing", ## status This is repeated for each review
"No new legal issues per client report. "
],
[
"09/06/2014",
"01/05/2015",
"Ongoing",
"Client reported attending court for probation violation due to lack of payment. "
],
[
"09/06/2014",
"10/23/2014",
"Achieved",
"No new legal issues per client report. "
],
[
"09/06/2014",
"08/28/2014",
"Ongoing",
"No new legal issues per client report. "
],
[
"09/06/2014",
"07/30/2014",
"Ongoing",
"No legal charges per client report. "
],
[
"09/06/2014",
"06/05/2014",
"Ongoing",
"No new legal issues per client report. "
],
[
"09/06/2014",
"05/08/2014",
"Ongoing",
"No legal issues per client report. "
],
[
"09/06/2014",
"03/10/2014",
"Ongoing",
"Initial creation."
],
[
"tDate",
"rDate",
"Sts",
"Just"
]
]
],
[
[
"Substance Abuse", ## goal 2
" Copy: Successfully complete Intensive Outpatient Treatment (IOPT) program.",
" Refer client to IOPT Treatment program.",
" Frequency: At least once per month and as needed.",
" Care Manager and He will discuss progress as it relates to Outpatient Treatment.",
"Ted Therapist, CM\n\nClient \n\nTreatment Provider ",
104,
"03/10/2014" ## date created
],
[
[
"09/06/2014",
"03/19/2015",
"Discontinued", ## status
"Client reported 2/2015 as last date of use for cannabis. "
],
[
"09/06/2014",
"01/05/2015",
"Ongoing",
"Client denies subtance use over last couple months. "
],
[
"09/06/2014",
"10/23/2014",
"Ongoing",
"Client reports smoking cannabis 3 weeks ago due to family issues. He reports testing
positive at office meeting today. "
],
[
"09/06/2014",
"08/28/2014",
"Ongoing",
"No substance use per client report. "
],
[
"09/06/2014",
"07/30/2014",
"Ongoing",
"Client reports using marijana one month ago. "
],
[
"09/06/2014",
"06/05/2014",
"Ongoing",
"No substance use per client report. "
],
[
"09/06/2014",
"05/08/2014",
"Ongoing",
"No substance use per client report. "
],
[
"09/06/2014",
"03/10/2014",
"Ongoing",
"Initial creation."
],
[
"tDate",
"rDate",
"Sts",
"Just"
]
]
],
[
[
"Domain",
"Goal",
"Interventions",
"Frequency",
"Progress",
"Clinician",
"Base",
"Created"
],
[
"tDate",
"rDate",
"Sts",
"Just"
]
]
],
"Strengths": "Supportive family \ngood person \nhard worker",
"zzud_group": "E8163487-E654-4D07-A13C-C2CAEB4EDB24",
"noReviewDate": "09/06/2014",
"newPlan": "",
"is_saved": "33E8C6A4-9DC8-451F-897F-EE7597619FFF",
"name_last": "Rabbit"
}
答案 0 :(得分：1)
将您的JSON数据存储在 jsonb 列中。强制完整性，需要较少的磁盘空间并允许索引。此外，对于大多数有用的操作（例如您的查询），无论如何都需要转换为json或jsonb。
不要使用 ，这是没用的。SIMILAR TO
而是使用 ISO 8601日期格式，这与任何区域设置或DateStyle设置无关。 （'2015-04-15'，而不是 '04 / 15/2015'）
使用显式JOIN语法和表别名来获取更易读的SQL代码。
您的查询已重新格式化，并使用LIKE而不是SIMILAR TO，但其他方面尚未更改（仍无效）：
SELECT c.name_lastfirst_cs
, c.client_id
, n.date_service
, s.code
, s.name
, n.data
, n.zrud_templat
, t.staff_name_cs
JOIN public.services s
JOIN public.notes n ON n.zrud_service = s.zzud_service
FROM public.clients c ON c.zzud_client = n.zrud_client
JOIN public.staff t ON t.zzud_staff = n.zrud_staff
WHERE s.code = '000502'
AND n.date_service BETWEEN '2014-07-01' AND '2015-04-15'
AND n.data LIKE '%Legal%' -- problematic predicates
AND n.data LIKE '%Achieved%'
AND n.data LIKE '%Substance Abuse%'
AND n.data LIKE '%Achieved%'
ORDER BY c.name_lastfirst_cs;
对事物的确切定义很有可能。您当前的查询或解释并未说明与完全匹配的内容。除非您知道数组中的位置，否则嵌套的JSON数组很难处理。并非不可能，但很难。
假设 jsonb 数据类型。如果尚未更改数据类型，请更改数据类型：
ALTER TABLE notes ALTER COLUMN data TYPE jsonb USING data::jsonb;
如果位置没有改变，"Legal"始终是嵌套数组第3级中的第一个元素，"Substance Abuse"是第二个，最新条目来自首先，数组中的位置是固定的。 IOW：通过您提供的示例，这将起作用：
AND n.data->'GoalsArray'->0->0->>0 = 'Legal'
AND n.data->'GoalsArray'->0->1->0->>2 = 'Achieved'
AND n.data->'GoalsArray'->1->0->>0 = 'Substance Abuse'
AND n.data->'GoalsArray'->1->1->0->>2 = 'Achieved'
其他您必须使用jsonb_array_elements()取消JSON数组（递归）并编写更复杂的查询。 WITH ORDINALITY对于将未使用的元素附加唯一数字非常有用：
如何取消JSON数组：
您可以使用索引：